Overview

The calculate statement is used to perform arithmetic operations using literal and column values.

Syntax

calculate columnResultColas source operation source

where source is either of columncolName or valueliteral_value

and operation is one of the characters + - * / % for addition, subtraction, multiplication, division and modulo respectively.

Examples: calculate column ResultCol as column Amount * value 1.2 calculate column Net as column total - column cogs calculate column constant_7 as value 3.5 + value 3.5

Details

The ResultCol parameter is the name of the column that will hold the results. This column may or may not exist (if necessary it will be created automatically).

Both of the two source parameters can specify a literal value, or the name of a column containing the value to use when performing the calculation.

• A literal value is specified using valueN where N is the literal number required

• A column name is specified using columncolName where ColName is the name of the column containing the values required

Additional notes:

• Any blank or non-numeric values in a source column will be treated as 0

• An attempt to divide by zero will result in 0

• When performing a modulo operation, the two source values are rounded to the nearest integer first

• If the result column already exists then if option overwrite is set to no, only blank cells in the result column will be updated.

Examples

• Add 1.5 to the values in the Rate column:

calculate column Rate as column Rate + value 1.5

• Multiply the values in the Rate column by those in the Quantity column

• Store the result in a new column called Charge

calculate column Charge as column Rate * column Quantity

Syntax

Within the individual reference articles for each statement, the syntax is described using the following conventions:

• bold for keywords

• italics for arguments

• Square brackets for optional keywords and arguments [likethis]

• Vertical pipe for alternative keyword options just|exactly as shown

• Ellipses for a variable length list of arguments: Column1 ... ColumnN

Refer to the page for more information regarding datasets, fully qualified column names and related information.

Reference

The following statements (in alphabetical order) are supported by Transcript:

Overview

The capitalise (the spelling capitalize is also supported) statement is used to modify the name of a column and/or the values in a column such that the first character is a capital letter and the remaining characters are lower case.

Syntax

capitalise values|heading [and values|heading] in column|columnsColName1 [... ColNameN]

Details

The heading and values keywords refer to the name of an existing column and the values in each row for that column respectively.

Only the first character in the column name or value is converted to upper case. If this character is not a letter, then the statement will have no effect on it. For example applying the statement:

will have no effect on the column name as the first character is an underscore. However, applying the same statement to a column called _VMName would result in a new name of _vmname as after attempting to make the first character a capital (which in the case of the underscore has no effect), the remaining characters are converted to lower case.

Any number of column names may be specified, and any of these may or may not be fully qualified. When applied to values in a column, blank values are ignored.

Examples

Overview

The create statement is used to add one more more new columns to an existing DSET.

Syntax

create columnNewColumnName[valueValue]

create columns fromColumnName[usingValueColumnName]

create mergedcolumnNewColumn[separatorsep]from [stringliteral] Column [/regex/] [Column [/regex/]|stringliteral]

Details

Explicit single column creation

Syntax

create columnNewColumnName[valueValue]

Details

This statement is used to create a new column called NewColumnName. The NewColumnName argument may be a column name, in which case the new column will be created in the DSET specified as part of that name.

Note: If no DSET has been explicitly defined using the statement then the DSET created by the first or statement in the Transcript task is automatically set as the default DSET.

A column called NewColumnName must not already exist in the DSET. If NewColumnName contains dots then they will be converted into underscores.

The new column will be created with no values in any cells, unless the optional value *Value* portion of the statement is present, in which case all the cells in the new column will be set to Value.

Examples

Create a new empty column called Cost in the default DSET: create column Cost

Create a new column called Cost with a value of 1.0 in every row of the default DSET: create column Cost value 1.0

Create a new column called Cost with a value of 1.0 in every row of the DSET custom.charges: create column custom.charges.Cost value 1.0

Automated single/multiple column creation

Syntax

create columns fromColumnName[usingValueColumnName]

Details

This statement is used to create multiple columns in a single operation. As is the case for create columns above, if the using ValueColumnName portion of the statement is not present, then all newly created columns will have no values in any cells.

Given this example dataset:

The statement create columns from ServiceName using Count will create the result shown below:

The names of the new columns to create are derived from the contents of the cells in the column called ColumnName, and the values (if opted for) are derived from the contents of the cells in the column called ValueColumnName. Duplicates are ignored. If all the cells in ColumnName have the same contents, then only a single new column will be created. To illustrate this, consider the following:

When applied to the data above, the statement create columns from ServiceName will produce the following result (note that only a single column called Small_VM is created, and that empty cells are represented with a separator character, which in the case of the below is a comma):

If opting to set the values in the new columns, then for each row the value in ValueColumnName will be copied into the column whose name matches ColumnName. When applied to the same original data, the statement create columns from ServiceName using Quantity will produce the following result:

When using create columns the new columns are always created in the default DSET. This means that when no values are being set, it is possible to specify a different DSET for ColumnName. If the default DSET is Azure.usage, then the statement create columns from custom.data.Services will derive the names of the new columns from the cell contents in the Services column in the custom.data DSET.

This is only possible in the absence of the using ValueColumnName option. When values are to be set, both the ColumnName and ValueColumnName arguments must belong to the default DSET.

Example

The following transcript task will import the datasets Azure.usage and system/extracted/Services.csv, and create new (empty) columns in Azure.usage whose names are taken from the values in the column ServiceDefinitions in Services.csv.

Merging column values to create a new column

Syntax

create mergedcolumnNewColumn[separatorsep]from [stringliteral] Column [/regex/] [ ... Column [/regex/]|stringliteral]

Details

This form of the statement is used to generate a new column containing values derived from those in one or more existing columns (termed source columns). The parameters are as follows:

If a regex is specified then it must contain a subgroup enclosed in parentheses. The portion of the text in the source column matched by this subgroup will be extracted and used in place of the full column value.

If a regex is not specified, then the entire value in the source column will be used.

Options

By default the value extracted from a source column will be blank in the following two cases:

• There is a blank value in a source column

• No match for a regular expression is found in the value of a source column

In such cases the merged result will simply omit the contribution from the source column(s) in question. If all the source columns yield a blank result then the final merged result will also be blank.

option merge_blank = some_text_here

This option will use the string some_text_here in place of any blank source column value.

option merge_nomatch = some_text_here

This option will use the string some_text_here if the result of applying the regular expression to a column value returns no matches.

Specifying the literal string <blank> as the merge_blank or merge_nomatch value will reset the option such that the default behaviour is re-activated.

Examples

Given the following dataset:

The following examples illustrate some uses of the create mergedcolumn statement:

Example 1

Example 2

If no regular expression is specified then the values in the source column will be used in their entirety:

Example 3

Let us add a new row to the sample dataset which has a non-compliant value for the user_id:

By default a non-matching value will result in a blank component of the merged result:

In this case, the resulting key for John has no separator characters in it. We can force a default value for the missing user_id portion as follows:

Overview

The aggregate statement is used to reduce the number of rows in a DSET while preserving required information within them

Syntax

aggregate [counter_columncolname] [dset.id][notime|daily] [offsetoffset][nudge] [default_functionfunction] colname function [... colname function]

Details

The aggregate statement is a powerful tool for reducing the number of rows in a DSET. Aggregation is based on the concept of matching rows. Any two rows that match may be merged into a single row which selectively retains information from both of the original rows. Any further rows that match may also be merged into the same result row.

A quick introduction

A match is determined by comparing all the columns which have a function of match associated with them (further information regarding this can be found below). If all the column values match, then the rows are merged.

Merging involves examining all the columns in the data that were not used in the matching process. For each of those columns, it applies a function to the values in the two rows and updates the result row with the computed result of that function. For a full list of functions, please refer to the table further down in this article.

To illustrate this consider the following two row dataset:

If we don't care about the colour value in the above records, we can combine them together. We do care about the quantity however, so we'll add the two values together to get the final result.

The statement to do this is:

aggregate notime id match location match quantity sum

• id match means that the values in the id columns must be the same

• location match means that the values in the location columns must be the same

• quantity sum means that the resulting value should be the sum of the two existing values

• by default, a function of first is applied to the columns, such that the original row retains its value

Applying these rules to the above example we get the following single result record:

Parameters

The aggregate statement accepts a range of parameters as summarised in the table below:

If two records are deemed suitable for merging then the function determines the resulting value in each column. The available functions are as follows:

The counter column

Each of the rows in a DSET that has been aggregated may be the result of merging multiple rows that existed prior to aggregation. The aggregate statement therefore creates a new column (or overwrites an existing column if told to) which for each row contains a count of the number of rows that were merged to create that row.

By default this column is called EXIVITY_AGGR_COUNT but an alternative name may be specified using the optional counter_column parameter.

For example the following statement will perform the aggregation and place the counts for each row into a column called merged_row_count

If the column specified by counter_column does not exist then it will be created.

If the column does exist then the contents will be overwritten.

Non time-sensitive aggregation

When the notime parameter is specified, the aggregation process treats any columns flagged as start and end times in the data as data columns, not timestamp columns.

In this case when comparing two rows to see if they can be merged, the aggregation function simply checks to see if all the columns with a function of match are the same, and if they are the two rows are merged into one by applying the appropriate function to each column in turn.

De-duplication

The following illustrates the aggregate statement being used to remove duplicate rows from a DSET:

The analysis of the statement above is as follows:

• notime - we are not interested in timestamps

• default_function match - by default every column has to match before records can be aggregated

• subscription_id match - this is effectively redundant as the default_function is match but needs to be present because at least one pair of colname function parameters is required by the aggregate statement

The resulting DSET will have no duplicate data rows, as each group of rows whose column values were the same were collapsed into a single record.

Row reduction while preserving data

The example shown at the top of this article used the sum function to add up the two quantity values, resulting in the same total at the expense of being able to say which source record contributed which value to that total.

The sum function can therefore accurately reflect the values in a number of source rows, albeit with the above limitation. By using a function of sum, max or min, various columns can be processed by aggregate in a meaningful manner, depending on the specific use case.

Time-sensitive aggregation

When aggregating, columns containing start time and end time values in UNIX epoch format can be specified. Each record in the DSET therefore has start and end time markers defining the period of time that the usage in the record represents. As well as taking the start times and end times into account, time-sensitive aggregation can perform additinal manipulations on these start and end times.

A quick example

Consider the following CSV file called aggregate_test.csv:

It is possible to aggregate these into 3 output records with adjusted timestamps using the following Transcript task:

Resulting in:

As can be seen, for each unique combination of the values in the id,subscription-id and service columns, the start and end times have been adjusted as described above and the quantity column contains the sum of all the values in the original rows.

Further notes

The daily parameter to aggregate means that the START_TIME and END_TIME columns are now recognised as containing timestamps. When aggregating with the daily option, timestamps within the current dataDate are combined to result in an output record which has the earliest start time and the latest end time seen within the day.

Optionally, following daily an offset may be specified as follows:

aggregate aggr.test daily offset 2 id match subscription_id match quantity sum

In this case the start and end timestamps are adjusted by the number of hours specified after the word offset before aggregation is performed. This permits processing of data which has timestamps with timezone information in them, and which may start at 22:00:00 of the first day and end at 21:59:59 of the second day, as an offset can be applied to realign the records with the appropriate number of hours to compensate.

The nudge parameter shaves 1 second off end times before aggregating in order to avoid conflicts where hourly records start and end on 00:00:00 9the last second of the current hour is the same as the first second of the next hour)

Overview

The default statement is used to explicitly define the default DSET to use when specifying a column name as an argument in subsequent statements.

Syntax

default dsetsource.alias

Details

Given that multiple DSETS can be loaded at once, it is necessary to specify which DSET to operate on when performing actions such as creating a new column with create. A column name in a Transcript statement is assumed as belonging to the default DSET unless it is a fully qualified column name.

If there is no default statement in the Transcript, then the first CSV file imported via the import statement will automatically be designated as the default DSET.

The default statement can be used multiple times throughout a Transcript, in which case the default DSET will be whichever was specified by the last default statement executed.

Lastly, when executing a finish statement, unless otherwise specified, the default DSET will be used to populate the reporting database created as a result.

Examples

Set custom.datafile as the default DSET:

default dset custom.datafile

Overview

The delete statement is used to delete one or more columns or rows from one or more DSETs.

Syntax

delete columns [except]ColName1 [... ColNameN]

delete blankcolumns

delete rows

delete dsetdset_id

Details

Deleting columns

The list of column names following the initial delete columns statement may be used in one of two ways:

1. The columns listed will be deleted

2. All columns except those listed will be deleted

Which method to use is determined by the presence (or otherwise) of the except keyword:

delete columns One Two Three         # Delete the listed columns
delete columns except One Two Three  # Delete all but the listed columns

It is not possible to delete a column from a DSET which only contains a single column. When deleting a column, the memory used to index it, along with the memory used to store the contents of its cells is released. This may be useful in situations where memory is limited.

Deleting unwanted columns early in a Transcript task will also increase the the performance of many subsequent operations performed by the task.

The second keyword may be column instead of columns. Either way, the statement behaves identically.

Deleting blank columns

The delete blankcolumns statement will delete columns that only contain blank values from the default DSET.

Deleting rows

In order to delete rows from a DSET a local filter must be in effect. A local filter is created using the where statement.

When used within the body of a where statement, delete rows will delete all rows in the DSET associated with the local filter where the condition is true. For more information on conditions, please refer to the where article.

Deleting a DSET

The delete dset statement can be used to remove a DSET from memory. This may be useful in cases where a DSET is no longer required, for example after it has been used to enrich another DSET via the correlate statement.

Examples

Delete the column temp from the default DSET and the column Interim from the azure.Usage DSET: delete columns temp azure.Usage.Interim

Delete any columns in the default DSET that have no values in any rows: delete blankcolumns

Delete all rows where the VMID is 1234

where ([VMID] == 1234) {
    delete rows
}

Delete the DSET azure.rates delete dset azure.rates

Overview

The correlate statement is used to enrich the default DSET by adding new columns to it, and/or updating existing columns with useful values. The new column names are derived from other DSETs and the values in those columns are set using a lookup function based on the value in a key column shared between the DSETs.

Syntax

correlateColName1 [ ... ColNameN]usingKeyColumn[assumingassumeDSET][defaultDefaultValue]

Details

The ColName1 ... ColNameN arguments are column names that will be copied from their original DSETs and merged into the default DSET.

Column names must be fully qualified, unless the assuming parameter is used, in which case any column names that are not fully-qualified will be assumed to belong to the DSET specified by assumeDSET.

Source and Destination columns

Source columns are those from which a cell is to be copied when the KeyColumn matches. Destination columns are columns in the default DSET into which a cell will be copied. Destination column names are derived from the names of the source columns as follows:

• The source column is the argument in its original form, for example: Azure.usage.MeterName

• The destination column is the same argument, but with the DSET ID replaced with that of the default DSET. For example if the default DSET is Custom.Services then the destination column for the above would be Custom.Services.MeterName.

If a destination column name doesn't exist in the default DSET then a new column with that name will automatically be created.

The Key Column

The KeyColumn argument is a column name which must not be fully qualified and which must exist in the default DSET and all of the DSETs referenced by the ColNameN arguments.

Default values

The DefaultValue argument, if present, specifies the value to write into the destination column if there is no match for the KeyColumn. If the DefaultValue argument is not specified then any rows where there is no match will result in a blank cell in the destination column.

For each row in the default DSET, the source DSET is searched for a matching KeyColumn value, and if a match is found then the value in the source column is used to update the default DSET. The row of the first match found in the source DSET will be used.

Overwriting

When matching the KeyColumn values, the logic in the following table is evaluated against every row in the destination DSET.

✘ means no or disabled, ✔ means yes or enabled

Examples

Given two Datasets as follows, where the default DSET is MyData.Owners:

Dataset 'MyData.Owners'

owner,id
John,100
Tim,110
Fokke,120
Joost,130
Jon,140

Dataset 'Custom.Services'

service,description,id
Small_VM,Webserver,130
Medium_VM,App_Server,100
Large_VM,DB_Server,110
Medium_VM,Test_Server,120

The statement: correlate service description using id assuming Custom.Services

Will enrich the MyData.Owners Dataset such that it contains:

owner,id,service,description
John,100,Medium_VM,App_Server
Tim,110,Large_VM,DB_Server
Fokke,120,Medium_VM,Test_Server
Joost,130,Small_VM,Web_Server
Jon,140,,

The statement: correlate service description using id assuming Custom.Services default unknown

Will produce:

owner,id,service,description
John,100,Medium_VM,App_Server
Tim,110,Large_VM,DB_Server
Fokke,120,Medium_VM,Test_Server
Joost,130,Small_VM,Web_Server
Jon,140,unknown,unknown

Overview

The convert statement is used to convert values in a column from base-10 to base-16 or vice-versa.

Syntax

convertcolNameto decimal|hex from decimal|hex

Details

When converting values in a column, the following considerations apply:

• Values in the column are replaced with the converted values

• The colName argument must reference an existing column, and may optionally be fully qualified (else the column is assumed to be in the default DSET)

• If any values in the column are not valid numbers, they will be treated as 0

• Blank values are ignored

• The convert statement may be used in the body of a where statement

• If a value in colName contains a partially correct value such as 123xyz then it will be treated as a number up to the first invalid character, in this case resulting in a value of 123.

• The hex digits in the original value can be either upper or lower case

• The hex digits from A-F will be rendered in upper case in the converted output

• The convert statement only supports integer values (floating points will be treated as floored to the nearest integer)

Example

Overview

The export statement is used to snapshot the data in a DSET and write it to disk as a Dataset.

Syntax

exportsource.aliasasfilename

Details

The exported file will be created under <base_dir>/exported. The filename parameter may include a path as well as the filename to export, so long as it does not contain the substring ".." and is not an absolute path. If the path contains one or more directories that do not exist then they will be created.

Exporting usage

The source.alias argument is the DSET ID to export (see for more information on DSET IDs).

The export statement can be used at any point during this process to save a CSV (conforming to Dataset format) which snapshots the data in the DSET at that moment in time. This can be used for a number of purposes including:

• Examining the state of data part-way through processing for debugging purposes

• Creating custom exports for subsequent import into 3rd part systems

• Producing output CSV files (which may potentially contain merged data from multiple sources) for subsequent processing with another Transcript

Examples

The following Transcript will import a quoted CSV file, add a ProcessedFlag column to it with a value of 1 in each row, and save it back out again without quotes:

Overview

The append statement is used to append one to the end of another.

Syntax

appendsource_dset.idtodestination_dset.id

Details

If the source DSET has any column names not present in the destination DSET then additional columns are automatically created in the destination DSET. These additional columns will contain blank values by default.

If one or more column names are present in both DSETs then the columns copied from the source DSET may be re-ordered into the same order as that used by the destination DSET.

At the end of the operation, the destination DSET will contain all the data from both DSETs, and the source DSET is unchanged.

Example

Given the following DSETs:

DSET ID: example.data

DSET ID: example2.data

The statement append example2.data to example.data will result in the following destination DSET (example.data):

Overview

The finish statement creates a (RDF) from a DSET. The RDF can subsequently be used by the reporting engine.

Syntax

finish[dset.id]

Details

The finish statement is used to create an RDF from a DSET. Only a single DSET can be used to create an RDF, but multiple finish statements may be used within the same task file. If there is no dset.id parameter then the default DSET will be used.

The RDF created by finish will be saved as <BaseDir>\system\report\<yyyy>\<MM>\<dd>_source.alias.rdf

where:

• <yyyy> is the 4-digit year

• <MM> is the 2-digit month

• <dd> is the 2-digit day

• source.alias are the tags which form the DSET ID

Any existing RDF with the same name will be overwritten.

Examples

Create a Reporting Database file for the default DSET: finish

Create a Reporting Database file for the DSET Azure.usage finish Azure.usage

Overview

The event_to_usage statement generates new CSV file with usage records from START/STOP/UPDATE events in source .

Syntax

event_to_usage fromsource.aliastofile{ options }

See Details section for options description

Details

This statement produces usage records from events in source DSET. Three times of events are supported:

• START event - marks the start of consumption

• STOP event - marks the end of consumption

• UPDATE event - marks the change of consumption attributes, such as quantity

There are several situations when the usage record is created:

• from START to first matching STOP event

• from START to first matching UPDATE event

• from UPDATE to first matching STOP event

• from UPDATE to first matching UPDATE event

• from the beginning of the day to STOP event (if consumption started during previous days)

• from the beginning of the day to UPDATE event (if consumption started during previous days)

• from START event to the end of the day (if no matching STOP/UPDATE events found)

• from UPDATE event to the end of the day (if no matching STOP/UPDATE events found)

• for the whole day (if consumption started during previous days, and there was no STOP or UPDATE event during processing day)

Event B is considered matching to event A if happened after event B and has matching key fields.

Options

Several options control the behaviour of this statement:

Initial data load

If epoch_date option is specified, it is possible to perform initial data load - load running consumptions on specific date. Transcript performs following checks:

• there are no events loaded for the specified DSET

• processing date matches specified epoch_date

If any of these checks fail, Transcript stops with error.

Data integrity checks

It is very important to load events in correct order, therefore Transcript performs following checks:

• the data for the same day for the specified DSET cannot be loaded twice

• there cannot be gaps in processed dates (except for Initial data load)

If any of these checks fail, Transcript stops with error.

If there is a need to re-process data for specific day, event-related state in the database must be rolled back to the preceding date, and after data for all following days processed in correct order.

Examples

Only START and STOP events, single-column key:

START, STOP and UPDATE events, complex key:

Overview

The copy statement is used to copy rows from one DSET to another

Syntax

copy rows todset.id

move rows todset.id

Details

Both copy and move must be used within the body of a statement. Only rows that match the expression will be copied (or moved).

• The DSET from which rows will be copied or moved is automatically determined from the expression used by the where statement.

• The DSET to which rows will be copied or moved is determined by the dset.id parameter

The destination DSET may or may not exist. If it does not exist then it will be created. If it does exist then the following logic is applied:

• If the destination DSET has more columns than the source DSET then the new rows in the destination DSET will have blank values in the rightmost columns

• If the destination DSET has fewer columns than the source DSET then the destination DSET will be extended with enough new columns to accomodate the new rows. In this case, existing rows in the destination DSET will have blank values in the rightmost columns

If the destination DSET is extended to accomodate the source rows then the new (rightmost) columns will have the same names as the equivalent columns in the source DSET. In the event that this would cause a naming conflict with existing columns in the destination DSET, one or more new columns in the destination DSET will heve a suffix added to their name to ensure uniqueness. This suffix takes the form _N where N is a number starting at 2.

To illustrate this, if the source DSET has columns called subscription,name,address,hostname and the destination DSET has a single column called name then the resulting extended destination DSET would have columns called name,subscription,name_2,address,hostname.

Example

Overview

The normalise statement is used to update the values in a numerical column such that they are all positive, negative or inverted.

Syntax

normalise columncolNameas positive

normalise columncolNameas negative

normalise columncolNameas invert

normalise columncolNameas standard

Details

The normalise statement processes each value in the column called colName and applies the following logic based on the last argument shown above as follows:

In order to be considered a number, a value in the colName column must start with any of the characters +, -, . or 0 to 9 and may contain a single . character which is interpreted as a decimal point.

Additionally:

• Any numerical value in colName which starts with a +, . or decimal character is considered positive

• Any numerical value in colName which starts with a - character is considered negative

• When using standard the resulting conventional number will be accurate up to 14 decimal places

Example

The environment statement specifies the name of environment to use for resolving global variables.

Syntax

environment name

Details

The environment statement selects the predefined environment to use for global variable lookup. It is and error to specify the environment which is not defined in global database.

If no environment specified, default environment (the one specified as default in global database) is assumed.

Overview

The include statement is used to combine Transcript task files together

Syntax

includetaskfile

Details

The include statement is used to combine two or more task files into a single script prior to commencement of execution.

Before executing a task file, Transcript analyses the script and processes all the include statements. Wherever an include statement is encountered, the specified taskfile is read from disk and inserted into the script in place of the include statement itself.

Command placement

As with most statements, include must occupy a single line of its own in the script.

Additional notes

The taskfile argument is treated as relative to <basedir>/system/config/transcript/.

Multiple files may be included in a task file and any included file may include other files but a check is done to ensure that no infinite include loops are generated.

Example

Given the following two files in /system/config/transcript/ ...

# FILE 1: myscript.trs
option loglevel = DEBUGX

# import usage data from csp
import "system\extracted\AzureCSP\${dataDate}_csp_usage.csv" source CSP alias Usage
default dset csp.usage
rename column resource_id to meter_id

include import_customers.trs

option overwrite = no
set resource_subcategory to Generic
create column interval value individually
create mergedcolumn service_name separator " - " from resource_name resource_subcategory region
# etc ...

# FILE 2: import_customers.trs
import "system\extracted\AzureCSP\${dataDate}_csp_customers.csv" source CSP alias Customers
rename column CSP.Customers.ID to customer_id
# END FILE 2

Prior to execution the pre-processor will combine these into the single memory-resident Transcript task:

# FILE 1: myscript.trs
option loglevel = DEBUGX

# import usage data from csp
import "system\extracted\AzureCSP\${dataDate}_csp_usage.csv" source CSP alias Usage
default dset csp.usage
rename column resource_id to meter_id

# FILE 2: import_customers.trs
import "system\extracted\AzureCSP\${dataDate}_csp_customers.csv" source CSP alias Customers
rename column CSP.Customers.ID to customer_id
# END FILE 2

option overwrite = no
set resource_subcategory to Generic
create column interval value individually
create mergedcolumn service_name separator " - " from resource_name resource_subcategory region
# etc ...

Overview

The rename statement is used to change the name of an existing column in a DSET, or to change the source and/or alias of a DSET

Syntax

rename columnOldNametoNewName

rename dsetOldSource.OldAliastoNewSource.NewAlias

Details

Renaming a column

When renaming a column, OldName may be a fully qualified column name. If it is not fully qualified then the column OldName in the default DSET will be renamed.

The NewName argument must not be fully qualified.

Renaming a DSET

When renaming a DSET, NewSource.NewAlias must contain exactly one dot and must not be the name of an existing DSET. The rename takes immediate effect, thus any subsequent reference to the renamed DSET in the transcript task must use its new name.

Any pending service definitions referencing the renamed DSET will automatically be updated with the new name.

Examples

Renaming columns:

# Rename the column 'UnitCost' in the default DSET to 'Cost'
rename column UnitCost to Cost

# Rename the column 'UnitCost' in the DSET 'custom.prices' to 'CustomCost'
rename column custom.prices.UnitCost to CustomCost

Renaming a DSET:

# Usage file has yyyyMMdd in filename
import "system/extracted/AzureJuly/${dataDate}.ccr" source AzureJuly

# The resulting DSET will have an ID of 'AzureJuly.yyyyMMdd'
# where yyyyMMdd is the current dataDate

#
# Any additional processing would be done here
#

# Remove the yyyyMMddfrom the DSET ID
rename dset AzureJuly.${dataDate} to Azure.July

# Create 'Azure.July.rdf'
finish

Overview

The if statement is used to conditionally execute one or more statements

Syntax

if (conditional expression) {
    <statements ...>
} [else {
    <statements ...>
}]

Conditional Expressions

A conditional expression (hereafter referred to as simple an expression) is evaluated to provide a TRUE or FALSE result which in turn determines whether one or more statements are to be executed or not. The following are examples of a valid expression:

(${dataDate} == 20180801)

((${dataDate} >= 20180801) && ([hostname] == "templateVM"))

An expression used by the if statement may contain:

• Numeric and string literals

• Regular expressions

• Variables

• Operators

• Functions

Numeric and string literals

A literal is a specified value, such as 4.5 or "hostname". Literals may be numbers or strings (text).

If a literal is non-quoted then it will be treated as a number if it represents a valid decimal integer or floating point number (in either regular or scientific notation), else it will be treated as a string.

If a literal is quoted then it is always treated as a string, thus 3.1515926 is a number and "3.1415926" is a string.

Regular expressions

Regular expressions must be enclosed within forward slashes (/), and are assumed to be in ECMAScript format.

If present, a regular expression must be used on the right hand side of either an !~ or an =~ operator, and when evaluated it will be applied to the value on the left hand side of an operator, eg:

if (${dataDate} =~ /[0-9]{4}01/) {
    var first_day_of_month = yes
} else {
    var first_day_of_month = no
}

As the forward slash is used as a delimiter for the expression, any literal forward slashes required by the expression should be escaped with a back-slash: \/

Variables

Variables can be used within expressions, in which case they are replaced with their values. Once expanded, these values are treated as literals.

Operators

Operators are evaluated according to the operator precedence rules in the table below (where the highest precedence is evaluated first), unless parentheses are used to override them. Operators with the same precedence are evaluated from left to right.

Functions

A function is used to evaluate one or more arguments and return a result which is then taken into consideration when evaluating the overall truth of the expression.

Function calls start with a the character @ which is followed by the function name and a comma separated list of parenthesised parameters, for example @MIN(1, 2, 3) .

The following functions are supported by the if statement:

Numeric functions

MIN

@MIN(number, number [, number ...])

Return the smallest number from the specified list (requires at least 2 arguments)

Examples:

• @MIN(1,2) returns 1

• @MIN(1,2,-3) returns -3

• @MIN(1,2,"-1") returns -1 - string "-1" is converted to number -1

• @MIN(1,2,3/6) returns 0.5

• @MIN(1,2,"3/6") returns 1 - string "3/6" is converted to number 3, up to first invalid character

• @MIN(1,2,"zzz") returns 0 - string "zzz" is converted to number 0

MAX

@MAX(number, number [, number ...])

Return the largest number from the specified list (requires at least 2 arguments)

Examples:

• @MAX(1,2) returns 2

• @MAX(-1,-2,-3) returns -1

• @MAX(1,2,100/10) returns 10

ROUND

@ROUND(number [, digits])

Returns number rounded to digits decimal places. If the digits argument is not specified then the function will round to the nearest integer.

Examples:

• @ROUND(3.1415,3) returns 3.142

• @ROUND(3.1415,2) returns 3.14

• @ROUND(3.1415926536,6) returns 3.141593

• @ROUND(3.1415) returns 3

• @ROUND(2.71828) returns 3

String functions

CONCAT

@CONCAT(string1, string2 [, stringN ...])

This function will treat all its arguments as strings, concatenate them and return the result.

Examples:

• @CONCAT("the answer ", "is") returns the answer is

• @CONCAT("the answer ", "is", " 42") returns the answer is 42

• @CONCAT("the answer ", "is", " ", 42) returns the answer is 42

SUBSTR

@SUBSTR(string, start [, length])

Return a sub-string of string, starting from the character at position start and continuing until the end of the string end until the character at position length, whichever is shorter.

Examples:

• @SUBSTR("abcdef", 1) returns abcdef

• @SUBSTR("abcdef", 3) returns cdef

• @SUBSTR("abcdef", 3, 2) returns cd

• @SUBSTR("abcdef", 3, 64) returns cdef

STRLEN

@STRLEN(string)

Returns the length of its argument in bytes.

Examples:

• @STRLEN("foo") returns 3

• @STRLEN(@CONCAT("ab", "cd")) returns 4

• @STRLEN(1000000) returns 7 (the number 1000000 is treated as a string)

PAD

@PAD(width, value [, pad_char])

This function returns value, left-padded with pad_char (0 by default) up to specified width. If width is less than or equal to the width of value, no padding occurs.

Examples:

• @PAD(5, 123) returns 00123

• @PAD(5, 12345) returns 12345

• @PAD(1, 12345) returns 12345

• @PAD(5, top, Z) returns ZZtop

EXTRACT_BEFORE

@EXTRACT_BEFORE(string, pattern)

This function returns the substring of string that precedes the pattern. If pattern cannot be found in the string, or either string or pattern are empty, result of the function is empty string.

Examples:

• @EXTRACT_BEFORE("abcdef", "d") returns ab

• @EXTRACT_BEFORE("abcbc", "bc") returns a

• @EXTRACT_BEFORE("abcdef", "x") returns empty string

EXTRACT_AFTER

@EXTRACT_AFTER(string, pattern)

This function returns the substring of string that follows the pattern. If pattern cannot be found in the string, or either string or pattern are empty, result of the function is empty string.

Examples:

• @EXTRACT_AFTER("abcdef", "cd") returns ef

• @EXTRACT_AFTER("abcabc", "ab") returns cabc

• @EXTRACT_AFTER("abcdef", "abb") returns empty string

EXTRACT_XXX functions can be combined to extract the middle part of the string, for example @EXTRACT_AFTER(@EXTRACT_BEFORE("abcdef", "ef"), "ab") returns cd.

Date functions

CURDATE

@CURDATE([format])

Returns the current (actual) date in the timezone of the Exivity server. The format may be any valid combination of strftime specifiers. The default format is %Y%m%d which returns a date in yyyyMMdd format.

Examples (assuming run date is 1 July 2019, at 12:34:56):

• @CURDATE() returns 20190701

• @CURDATE(\"%d-%b-%y\") returns 01-Jul-19

• @CURDATE("%H:%M:%S") returns 12:34:56

• @CURDATE("%u") returns 1 (weekday - Monday)

• @CURDATE("%j") returns 182 (day of the year)

DATEADD

@DATEADD(date, days)

Adds a specified number of days to the given date, returning the result as a yyyyMMdd date.

Invalid dates are normalised, where possible (see example below):

Examples:

• @DATEADD(20180101, 31) returns 20180201

• @DATEADD(20180101, 1) returns 20180102

• @DATEADD(20171232, 1) returns 20180102 (the invalid date 20171232 is normalised to 20180101)

• @DATEADD(20180101, 365) returns 20190101

DATEDIFF

@DATEDIFF(end_date, start_date)

Returns the difference in days between two yyyyMMdd dates. A positive result means that date1 is later than date2. A negative result means that date2 is later than date1. A result of 0 means that the two dates are the same.

Invalid dates are normalised, when possible (see example below):

Examples:

• @DATEDIFF(20190101, 20180101) returns 365

• @DATEDIFF(20180201, 20180101) returns 31

• @DATEDIFF(20180102, 20180101) returns 1

• @DATEDIFF(20180101, 20180102) returns -1

• @DATEDIFF(20180101, 20180101) returns 0

• @DATEDIFF(20171232, 20180101) returns 0 (the invalid date 20171232 is normalised to 20180101)

DTADD

@DTADD(datetime, count [, unit])

This function adds count number of unit_s (DAYS by default) to the specified datetime value and return normalised result datetime value in YYYYMMDDhhmmss_ format.

Datetime can be in any of the following formats:

• YYYYMMDD

• YYYYMMDDhh

• YYYYMMDDhhmm

• YYYYMMDDhhmmss

All missing bits of datetime value assumed zeros.

Supported units are (both singular and plural spellings supported):

• YEAR

• MONTH

• DAY (default)

• HOUR

• MINUTE

• SECOND

Example

• @DTADD(20190701, 2) returns 20190703000000

• @DTADD(20190701, 2, HOURS) returns 20190701020000

• @DTADD(2019070112, 50, DAYS) returns 20190820120000

• @DTADD(20190701123456, 10, MONTH) returns 20200501123456

Transcript-specific functions

if (!@COLUMN_EXISTS("colName")) {
   The column colName does NOT exist
}

FILE_EXISTS

@FILE_EXISTS(filename)

Returns 1 if the file filename exists, else returns 0.

FILE_EMPTY

@FILE_EMPTY(filename)

In strict mode, this function returns 1 if the file filename exists and is empty. If the file does not exist, then this is considered an error.

In permissive mode, a non-existent file is considered equivalent to an existing empty file.

In either case, if the file exists and is not empty, the function returns 0

DSET_EXISTS

@DSET_EXISTS(dset.id)

Returns 1 if the specified DSET exists, else 0

DSET_EMPTY

In strict mode (option mode = strict), this function returns 1 if the specified DSET exists and is empty. If the DSET does not exist, then this is considered an error.

In permissive mode (option mode = permissive), a non-existent DSET is considered equivalent to an existing empty DSET.

In either case, if the DSET exists and is not empty, the function returns 0.

COLUMN_EXISTS

@COLUMN_EXISTS(column_name)

This function returns 1 if the specified column exists, else 0. The column name may be fully-qualified, but if it is not, then it is assumed to be in the default DSET.

DSET_ROWCOUNT

@DSET_ROWCOUNT(dset)

This function returns number of rows within specified DSET.

In permissive mode (option mode = permissive), a non-existent DSET is considered equivalent to an existing empty DSET, and zero is returned.

DSET_COLCOUNT

@DSET_COLCOUNT(dset)

This function returns number of columns within specified DSET.

In permissive mode (option mode = permissive), a non-existent DSET is considered equivalent to an existing empty DSET, and zero is returned.

Overview

The import statement is used to read a CSV file (which must conform to Dataset standards) from disk in order to create a DSET which can then be processed by subsequent Transcript statements.

Syntax

Import from CSV files

importfilenamefromsource[aliasalias][options { ... }]

importfilenamesourcecustom_source[aliasalias][options { ... }]

Import from CCR files

import filename.ccr source custom_source [aliasalias]

Import from database tables

import ACCOUNT sourcecustom_source[aliasalias]

import USAGE fordatefromdb_namesourcecustom_source[aliasalias]

Details

Options

Data imported from a CSV file can be filtered as it is being read in order to reduce post-import processing time and memory overhead.

The import filters are configured via the options parameter, which if present is imemdiately followed by one or more name = value pairs enclosed within braces, for example:

import system/extracted/example.csv source test alias data options {
    skip = 1
}

Multiple options may be specified and each option must be placed on separate line, for example:

import system/extracted/example.csv source test alias data options { 
    skip = 1
    omit = 4
}

The options supported by import are as follows:

The filter expression, if specified, is of identical format to that used by the where statement and it must reference at least one column name, enclosed within square brackets:

import system/extracted/example.csv source test alias data options { 
    # Only import rows where the vmType column is not the value "Template"
    filter ([vmType] != Template)
}

To specify a parameter or column name that contains spaces in an expression, use quotes within the square brackets as follows:

import system/extracted/example.csv source test alias data options { 
    filter (["service name"] =~ /.*D/)
}

File name / pattern

If a pattern is specified in the import options, then the filename parameter is treated as an ECMAScript-type regular expression, and all files matching that pattern are imported and appended to one another to result in a single DSET.

If using a regular expression, the import options are applied to all files matching that expression. All files must have the same structure (after any select/ignore options have been applied).

Database tables

Account data

To import all accounts from Exivity the following statement is used:

import ACCOUNT sourcecustom_source[aliasalias]

This functionality is intended for advanced use cases where it is necessary to correlate information about existing accounts into the new data being processed as part of the Transform step.

Usage data

To import usage data previously written to an RDF using finish the following statement is used:

import USAGE fordatefromdsetsourcecustom_source[aliasalias]

This functionality may be useful in the event that the original data retrieved from an API has been deleted or is otherwise unavailable.

The date must be inyyyyMMddformat. The RDF from which the usage will be imported is:

<basedir>/system/report/<year>/<month>/<day>_<dset>_usage.rdf.

To illustrate this in practice, the statement ...

import USAGE for 20180501 from azure.usage source test alias data

... will load the usage data from ...

<basedir>/system/report/2018/05/01_azure.usage_usage.rdf

... into a DSET calledtest.data.

Source and Alias tags

A file imported from disk contains one or more named columns which are used to create an index for subsequent processing of the data in that file. As multiple files may be imported it is necessary to use namespaces to distinguish between the DSETs created from the files (each imported file is converted into a memory-resident DSET before it can be processed further). This is accomplished through the use of source and alias tags. Each file imported is given a unique source and alias tag, meaning that any column in the data imported from that file can be uniquely identified using a combination of source.alias.column_name.

There are two main variations of the import statement which are as follows:

Automatic source tagging

importfilenamefromsource[aliasalias]

By convention, data retrieved by the USE extractor from external sources is located in a file called

<basedir>/system/extracted/<source>/<yyyy>/<MM>/dd>_.csv

where <yyyy> is the year, <MM> is the month and <dd> is the day of the current data date.

Typically, the <source> portion of that path will reflect the name or identity of the external system that the data was collected from.

By default the alias tag will be set to the filename, minus the `_ portion of that filename and _.csv` extension but an alias can be manually specified using the optional _alias parameter.

As an example, assuming the data date is 20170223, the statement:

import usage from Azure

will create a DSET called Azure.usage from the file /system/extracted/Azure/2017/02/23_usage.csv.

The statement:

import usage from Azure alias custom

will create a DSET called Azure.custom from the file

<basedir>/system/extracted/Azure/2017/02/23_usage.csv.

Manual source tagging

importfilenamesourcecustom_source[aliasalias]

This form of the statement will import filename and the source tag will be set to the value of the custom_source parameter.

By default the alias tag will be set to the filename minus the .csv extension but an alias can be manually specified using the optional alias parameter.

As an example, assuming the data date is 20170223, the statement:

import "system/extracted/Azure/${dataDate}.csv" source Azure alias usage

will create a DSET called Azure.usage fom the file

<basedir>/system/extracted/Azure/20170223.csv

Importing CCR files

When performing an import using manual tagging it is possible to import a Cloud Cruiser CCR file. This is done by specifying the .ccr filename extension as follows:

import system/extracted/myccrfile.ccr source MyData

Full information regarding CCR files is beyond the scope of this article. Cloud Cruiser documentation should be consulted if there is a requirement to know more about them.

In order to create a DSET from a CCR file, Transcript coverts the CCR file to a Dataset in memory as an interim step. The resulting DSET will have blank values in any given row for any heading where no dimension or measure in the CCR file existed that matched that heading. No distinction is made between dimensions and measures in the CCR file, they are imported as columns regardless.

When importing a CCR file, the quote and separator characters are always a qouble quote - " - and a comma - , - respectively. Fields in the CCR file may or may not be quoted. The conversion process will handle this automatically and any quotes at the start and end of a field will be removed.

Quote and Separator options

The fields in a dataset file may or may not be quoted and the separator character used to delineate fields can be any character apart from an ASCII NUL (0) value. It is therefore important to ensure that the correct options are set before importing a dataset in order to avoid unwanted side effects.

The options relating to import are quote and separator (or delimiter). By default, these are set to a double quote and a comma respectively.

If defined the quote character will be stripped from any fields beginning and ending with it. Any additional quote characters inside the outer quotes are preserved. Fields that do not have quotes around them will be imported correctly, unless they contain a quote character at only one end of the field.

If no quote character is defined, then quote characters are ignored during import, but any separator characters in the column headings will be converted to underscores.

Embeds

If the embed option is set, then a curly bracket in the data - { - will cause all characters (including newlines) up until the closing bracket - } - to be imported. Nested brackets are supported, although if there are an uneven number of brackets before the end of the line an error will be generated in the logfile and the task will fail.

Examples

Create a DSET called Azure.usage from the file /system/extracted/Azure/2017/02/23_usage.csv import usage from Azure

Create a DSET called Azure.custom from the file /system/extracted/Azure/2017/02/23_usage.csv import usage from Azure alias custom

Create a DSET called Azure.usage fom the file /system/extracted/Azure/20170223.csv import "system/extracted/Azure/${dataDate}.csv" source Azure alias usage

Create a DSET called Azure.usage from the file /system/extracted/Azure/20170223.csv import "system/extracted/Azure/${dataDate}.csv" source Azure alias usage

Overview

The split statement is used to create and/or update columns by splitting the textual value of an existing column into multiple parts.

Syntax

splitColNameusingsep

splitColNameusingsepretaining first[column_count]

splitColNameusingsepretaining last[column_count]

splitColNameusingsepretainingfirst_column_index[tolast_column_index]

Details

The ColName argument is the name of an existing column whose values are to be split and the sep argument (which must only be a single character in length) is the delimiter by which to split them.

For example a value one:two:three:four, if divided by a sep character of :, would result in the fields one, two, three and four.

New columns are created if necessary to contain the values resulting from the split. These additional columns are named ColName_split1 through ColName_splitN where N is the number of values resulting from the process.

If there is an existing column with a name conflicting with any of the columns that split creates then:

• • If there are blank values in the existing column they will be updated

  • If there are no blank values in the existing column then no changes will be made to it

Keeping only specific fields

If the retaining keyword is specified, split will discard one or more of the resulting columns automatically based on the specification that follows the keyword. The following specifications are supported:

Example

Given an input dataset of the form:

The statement ...

split ID using :

... will result in the dataset:

Using the same original dataset, the statement ...

split ID using : retaining 3 to 5

... will result in the dataset:

Overview

The option statement is used to set global parameters during the execution of a Transcript task.

Syntax

optionoption = setting

option noquote

Details

The option statement can be used multiple times within the same task script and always takes immediate effect. It is therefore possible (for example) to import a CSV file delimited with commas and quoted with single quotes, change the options and then export it with semicolon delimiters and double quotes.

The supported options are as follows:

Additional notes

Continue

option continue = yes|enabled

option continue = no|disabled

When executing a task file repeatedly against each day in a date range, by default Transcript will abort the whole run if a task failure occurs. In cases where this is undesirable, setting the continue option to enabled or yes (both work in exactly the same way) will change the behaviour such that if a task failure occurs then execution will resume with the next day in the range.

Delimiter / Separator

When specifying a quote or tab as the separator it is necessary to escape it in order to prevent Transcript from interpreting it as a meaningful character during the parsing of the task script. For example:

option delimiter = \t # Specify a literal TAB character
option delimiter = \" # Specify a literal quote

Services

This option may be specified as readonly , overwrite or update and influences the behaviour of the service and services statements as follows:

Execution mode

option mode = strict
option mode = permissive

Transcript supports two modes of execution for tasks:

• In strict mode, if an error is encountered at any point in the task, the error will be logged and execution will terminate

• In permissive mode, many errors that would otherwise have caused the task to fail will be logged, the statement that caused the error will be skipped and execution will continue from the next statement in the task.

Errors that can be handled in permissive mode are mainly syntax errors or those involving invalid parameters to Transcript statements. There are error conditions that can arise during the execution of a statement will cause the task to fail even in permissive mode.

Log levels

Transcript generates a considerable amount of logging information during the execution of a task. The loglevel option can be used to increase or decrease the level of detail written to the logfile. All logging levels must be specified in UPPER CASE. The following levels can be set:

The order of the logging levels in the table above is significant, in that for any given level, all levels above it in the table are also in effect. Therefore a logging level of WARN will result in log entries for ERROR, FATAL, and INTERNAL level events, as well as warnings.

Regardless of the logging level, some events will always create a logfile entry, for example the success or failure of a transcript task at the end of its execution.

Log mode

In order to minimise the effect on performance when logging, Transcript opens the logfile when it is first run and then holds it open until the completion of the task being executed. The logfile can be accessed in one of two modes:

The default is SAFE. It is not recommended that this be changed.

Examples

The following Transcript task will import a CSV file quoted with double quotes and delimited with commas and then export a copy with semicolons as delimiters and quoted with single quotes:

It also increases the logging level for the import statement

option quote = \"
option loglevel = DEBUGX
import usage from Azure
option loglevel = INFO
option separator = ;
option quote = '
export azure.Usage as c:\transcript\exported\azure_modified.csv

Overview

The replace statement is used to search for, remove, and optionally replace with new values, substrings of the values in a column.

Syntax

replacesubstringincolName[withreplacement]

Details

The replace statement will remove or replace all occurrences of substring in the values of a column. The parameters are as follows:

The colName argument may or may not be fully qualified, but must reference an existing column.

If the optional replacement string is provided then any occurrences of substring will be substituted with the specified value.

If the replacement string is not provided then all occurrences of substring will be removed from the values in colName.

Example

Given the following sample data:

PayerAccountName,TaxationAddress,ProductCode,ProductName,ItemDescription
John Doe,"123 Big St, Largetown, RH15 0HZ, United Kingdom",AWSDataTransfer,AWS Data Transfer,$0.00 per GB - EU (Germany) data transfer from EU (Ireland)
John Doe,"123 Big St, Largetown, RH15 0HZ, United Kingdom",AmazonS3,Amazon Simple Storage Service,$0.0245 per GB - first 50 TB / month of storage used
John Doe,"123 Big St, Largetown, RH15 0HZ, United Kingdom",AWSDataTransfer,AWS Data Transfer,$0.00 per GB - EU (Germany) data transfer from US West (Northern California)
John Doe,"123 Big St, Largetown, RH15 0HZ, United Kingdom",AWSDataTransfer,AWS Data Transfer,$0.090 per GB - first 10 TB / month data transfer out beyond the global free tier
John Doe,"123 Big St, Largetown, RH15 0HZ, United Kingdom",AWSDataTransfer,AWS Data Transfer,$0.00 per GB - EU (Germany) data transfer from US East (Northern Virginia)

The following script ...

import "system\extracted\example.csv" source test alias data
replace "John Doe" in PayerAccountName with "Finance Dept"
replace "RH15 0HZ, " in TaxationAddress
replace "United Kingdom" in TaxationAddress with UK
replace "AWS" in ProductCode

where ([ProductCode] == "AmazonS3") {
    replace "Amazon" in ProductCode
}

replace "Amazon " in ProductName
replace "AWS " in ProductName
replace "transfer from" in ItemDescription with -

... will produce the following output data.

"PayerAccountName","TaxationAddress","ProductCode","ProductName","ItemDescription"
"Finance Dept","123 Big St, Largetown, UK","DataTransfer","Data Transfer","$0.00 per GB - EU (Germany) data - EU (Ireland)"
"Finance Dept","123 Big St, Largetown, UK","S3","Simple Storage Service","$0.0245 per GB - first 50 TB / month of storage used"
"Finance Dept","123 Big St, Largetown, UK","DataTransfer","Data Transfer","$0.00 per GB - EU (Germany) data - US West (Northern California)"
"Finance Dept","123 Big St, Largetown, UK","DataTransfer","Data Transfer","$0.090 per GB - first 10 TB / month data transfer out beyond the global free tier"
"Finance Dept","123 Big St, Largetown, UK","DataTransfer","Data Transfer","$0.00 per GB - EU (Germany) data - US East (Northern Virginia)"

Overview

The lowercase statement is used to modify the name of, and/or the values in, a column such that any upper case characters are replaced with their lower case equivalent.

Syntax

lowercase heading|values [and heading|values] in columnColName1 [... ColNameN]

Details

The heading and values keywords determine the scope of the processing to be applied. The columns named in ColName1 ... ColNameN may or may not be fully qualified. If a column name is not fully qualified then the default DSET will be assumed.

Only upper case characters in a column name or value are converted. Numbers, punctuation and other symbols are ignored. Any blank values in a column are ignored.

Examples

The terminate statement will exit the transcript task immediately.

Syntax

terminate

Details

Normally a transformation script will finish execution when an error is encountered or when the end of the script file is reached, whichever comes first.

When the terminate statement is encountered, the script will finish at that point. No statements after the terminate statement will be executed.

Example

Overview

The timecolumns statement is used to set the start time and end time columns in a DSET

Syntax

timecolumnsstart_time_col end_time_col

timecolumns clear

Details

The usage data stored in a DSET may or may not be time sensitive. By default it is not, and every record is treated as representing usage for the entire day. In many cases however, the usage data contains start and end times for each record which define the exact time period within the day that the record is valid for.

If the usage data contains start and end times that are required for functions such as aggregation or reporting, the column(s) containing those times need to be marked such that they can be identified further on in the processing pipeline. This marking is done using the timecolumns statement.

The timecolumns statement does not perform any validation of the values in either of the columns it is flagging. This is by design, as it may be that the values in the columns will be updated by subsequent statements.

If the timecolumns statement is executed more than once, then only the columns named by the latest execution of the statement will be flagged. It is not possible to have more than one start time and one end time column.

Both the start_time_col and end_time_col parameters may be fully qualified column names, but they must both belong to the same DSET.

Clearing the flagged timestamp columns

To clear both the start and end time columns, thus restoring the default DSET to treating each record as spanning the entire day, the statement timecolumns clear may be used.

This can be useful in the following use case:

• The DSET is loaded and timestamp columns are created

• The timestamp columns are cleared

• Further processing is done on the DSET as required

Example

Overview

The set statement is used to write a specified value into all the cells in any given column, or to copy values from one column to another.

Syntax

setColNametoValue

setColNameasSrcColName

setColName=Expression

Details

If the keyword word to is used then column ColName will be updated using the constant value specified in Value.

If the word as is used then the values in the column SrcColName will be copied into column ColName.

If = is used then the expression is evaluated and the column values are set to the result. For a complete list of functions that can be used in an expression please refer to the in the article about the statement

The ColName argument must identify an existing column. The modified column will be in the default DSET unless it is a column name.

If the overwrite is disabled, then only cells with blank values will be updated with the new value.

Examples

Set the column called Flag in the default DSET to a value of 0:

Set the column called Flag to be the value in the column status:

Set the column called DiskTier to the 4th character in the existing value of that column

set DiskTier = @SUBSTR([DiskTier],4,1)

Set the column called Flag in the DSET custom.dataset to a value of Pending:

Set the column Flag to free if the rate column contains a value of 0:

Fill any blank values in the column username in the DSET Azure.usage to Unknown :

Overview

The round statement is used to ensure that numeric values in a column are whole multiples of a specified number.

Syntax

roundcolName [direction][to nearestvalue]

Details

The round statement is used to round numbers to the nearest multiple of any integer or floating point number.

The parameters supported by round are as follows:

The simplest form of the statement is round colName. This will use the defaults shown in the table above such that values are rounded to the next highest integer. Alternatively, the statement round colName down will round down to the nearest integer.

If the value argument is provided then the numbers in colName will be rounded up or down to the nearest multiple of value. For example the statement ...

round Quantity up to nearest 5

... will round the numbers in the column Quantity up to the next multiple of 5. Thus any number in Quantity higher than 10 and less than 15 will be rounded to 15.

Additional notes

The round statement observes the following rules:

• Non-numeric values in colName are ignored

• Blank values, or a value of 0 in colName are ignored

• Numbers in colName that are already a whole multiple of value will not be modified

Examples

Overview

The services statement is used to create or modify multiple services based on the data in a .

Syntax

services{ param1 = value [ ... paramN = value] }

Example:

Details

Summary

How column names are used

For many of the parameters to the services statement there are two ways of using a column name:

1. The values in the column are extracted from the usage data and those values are embedded as literals into the service definition.

2. The column name itself is used in the service definition such that the reporting engine dynamically determines the values to use for any given day when generating report data

Using the second method requires only a single rate revision which identifies by name the column(s) containing rate and/or COGS information. When a report is run, the charge engine then obtains the correct rate information for any given day from the data in those named columns.

Parameter table

The parameters supported by the services statement are summarised in the following table. The Type column in the table indicates the way the column name is used as described in How column names are used above. Additional information about each parameter can be found below the summary table itself.

Parameter details

usages_col

The usages_col parameter is the name of a column containing service keys. A service will be created for each distinct value in this column, and these values will be used as the service keys.

service_type

In order to calculate the charges associated with a service it is necessary to know the number of units of that service that were consumed. Exivity supports two methods of retrieving the units of consumption from usage data and the service_type determines which of these is applied.

Any given service may use one or other of these methods, which are as follows:

• Manual services: service_type = MANUAL

• Automatic services: service_type = AUTOMATIC

Manual services

Manual services require that the units of consumption for each service named in the usages_col column are stored in separate columns whose names correlate to the service keys themselves.

To illustrate this, consider the following fragment of usage data:

In this case, the service_name column contains a list of service keys and for each of those service keys there is a corresponding column containing the units of consumption for that service. Thus in the above example we can see that there are two services, "Small VM" and "Large VM" and that the units of consumption for each of these services are in the columns of the same name.

The more manual services that are represented in the data, the more columns are required.

Automatic services

Automatic services require that the units of consumption for each service named in the usages_col column are stored in the column named by the consumption_col parameter. To represent the same information as that shown in the example above, the following would be used:

It can be seen that any number of automatic services, along with their consumption figures, can be represented using only two columns of data.

consumption_col

The consumption_col parameter is only required when creating automatic services and determines the column containing the units of consumption for each service as described above.

instance_col

It is not enough to know only the units of consumption for each service, as this merely provides the total consumption for each service across the entire usage. In the examples above, for example, the "Large VM" service has 10 units of consumption but using that information alone there is no way to know if this represents one instance of a VM used 10 times, 10 instances of VMs used once each, or something in between.

The instance_col parameter is therefore required to tell the difference. Typically this will be a unique identifier which groups the units of consumption into 'buckets'. In the case of a VM this may be a VM ID which remains constant throughout the life of a VM in the cloud.

To illustrate this, we can supplement the example usage fragment used previously with additional information to use as the instance_col as follows:

By specifying instance_col = vmid we can now see that the usage represents:

• 5 instances of a single Small VM with an ID of 444

• 6 instances of a Large VM with an ID of 555

• 4 instances of a Large VM with an ID of 666

description_col

If specified, the description_col denotes a column containing a friendly description for reports to identify the service with.

Typically in cloud usage data, services are identified using unique IDs (referred to as keys in Exivity) which are often non-meaningful to human eyes, so Exivity supports a 'friendly' description for each service for display purposes when generating a report.

For example description_col = description may be used in conjunction with the following data to map the service_id to a friendly name:

It is not mandatory to provide a description_col parameter. If one is not supplied then the description will be set to a duplicate of the service key (as derived via the usages_col parameter).

category

By default Exivity will group services on reports according to their category. Using categories ensures that charges for services that fall within the same category wil appear in a contiguous block.

The category parameter specifies the category to which all services created will be assigned, thus specifying category = "Virtual Machines" might be appropriate for the example data used so far in this article.

category_col

Usage data normally contains information about a range of services of different types such as Virtual Machines, Storage, Networking and so on. By referencing a column in the usage data which identifies the correct category for each service, multiple categories will be created and each service assigned to the correct category by the services statement.

To illustrate this, let us extend the sample data as follows:

By specifying category_col = category each service will now be associated with the correct category.

interval

The interval parameter is used to specify a literal interval for all the services created by the services statement.

The interval parameter may be any of:

• individually

• daily

• monthly

If the interval parameter is not specified, then a default interval of monthly will be used.

interval_col

In the event that different services in the usages_col require different charge intervals, a column name containing the interval to use may be specified using the interval_col column as follows:

By specifying interval_col = interval each service in the above usage data will be assigned the correct charge interval.

model

The model parameter is used to enable proration for monthly services. Either of unprorated or prorated may be specified.

If no model is specified, then a value of unprorated will be used by default.

model_col

In the event that different services in the consumptions_col require different proration settings, the model_col parameter can be used to specify which column contains the proration setting for each service.

By specifying model_col = model, each service in the above usage data will be assigned the correct proration model.

charge_model

If specified as peak then the charge for any monthly services created will be calculated based on the day with the highest charge in the month.

If specified as average then the charge for any monthly services created will be calculated as the average unit price (for days that have usage only) multiplied by the average quantity (days with no usage will be treated as if they had a quantity of 0).

If specified as last_day then the charge for any monthly services created will be calculated based on the last day of the month.

If specified as day_xxx (where xxx is a number in 1-28 range) then the charge for any monthly services created will be calculated based on the specified day of the month.

charge_model_col

In the event that different services in the consumptions_col require different charge models, the charge_model_col parameter can be used to specify which column contains the charge_model setting for each service. For the example data below charge_model_col would be set to chargetype:

unit_label

The unit_label parameter is used by reports to provide a meaningful description of the units of consumption associated with a service. A virtual machine may have a unit label of Virtual Machines, but storage-related services may have a unit label of Gb for example.

If the unit_label parameter is not specified then a default lavel of Units will be used.

unit_label_col

In cases where the services contained in the usages_col column collectively require more than one unit label, the unit_label_col parameter can be used to identify a column in the usage data which contains an appropriate label for each service.

For example unit_label_col = label can be used to associate an appropriate label using the data below:

rate_col

The rate_col parameter is used to determine the column in the usage data which contains the unit rates for the service definitions created by the services statement.

As each service definition is created, an initial rate revision is also created which contains the column named by the rate_col parameter. When a report is run, for each day in the reporting range the unit rate for that day will be determined by whatever value is in the column named by the rate_col parameter in the usage data.

This means that only a single rate revision is required, even if the actual value in the rate_col column is different from day to day.

set_rate_using

The set_rate_using parameter is also used to determine the unit rate for each service. This differs from the rate_col parameter in that the values in the column named by set_rate_using are consulted when the service is created, and the literal values in that column are used to populate the initial rate revision.

This means that the unit cost is hard-coded into the rate revision and will apply indefinitely, or until such time as a new rate revision takes effect (see effective_date for more details)

cogs_col

The cogs_col parameter is used to determine the column in the usage data which contains the COGS rate associated with the service definitions created by the services statement.

As each service definition is created, an initial rate revision is also created which contains the column named by the cogs_col parameter. When a report is run, for each day in the reporting range the COGS rate for that day will be determined by whatever value is in the column named by the cogs_col parameter in the usage data.

If a monthly service has different COGS rates for different days in the month, then whichever results in the highest charge will be used.

This means that only a single rate revision is required, even if the actual value in the cogs_col column is different from day to day.

set_cogs_using

The set_cogs_using parameter is also used to determine the COGS rate for each service. This differs from the cogs_col parameter in that the values in the column named by set_cogs_using are consulted when the service is created, and the literal values in that column are used to populate the initial rate revision.

This means that the COGS rate is hard-coded into the rate revision and will apply indefinitely, or until such time as a new rate revision takes effect (see effective_date for more details)

set_min_commit_using

The set_min_commit_using parameter is used to set the minimum commit value in the initial rate revision for each service.

The values in the column identified by set_min_commit_using are extracted from the usage data and used as numeric literals in the revision.

effective_date

When creating the initial rate revision for a service, the value specified by the effective_date parameter is interpreted as a yyyyMMdd value to determine the date from which the revision should be applied.

effective_date_col

If there is a column in the usage data containing yyyyMMdd values representing the desired effective date for the initial revision of each service, The effective_date_col parameter may be used to extract the values from this column and set the effective date for each service accordingly.

Examples