When to use loops

You can create a pipeline (nested Flow) where some of the steps (inner Flows) are configured to be executed in a loop.

Typical use cases for executing Flows in a loop:

• Extract data from the HTTP endpoints with parameters. The parameters can be provided at runtime by using the loop.
• Use loop to implement API pagination. 
• Configure the transformations and Connections by reading configuration parameters from the database. 
• Configure Source query dynamically. 
• Dynamically setting up the credentials for the Connections.
• Executing Flow a fixed number of times.

Also, read about executing Flows conditionally. 

Loop types

Script Loop. Execute Flow while JavaScript or Python expression returns not null

Configure Script loop

Step 1. Create a nested Flow and add steps (inner Flows).

Step 2. Select the Flow to be executed in a loop and go to Flow parameters.

Step 3. Specify the Loop Type as Script.

Step 4. Add JavaScript or Python code in Loop Script.

Loop Script

The Flow continues to run in a loop while the JavaScript or Python expressions returns a not null value.

Example:

// this flow will be executed while the value of the "NEXT_MARKER" property is not null

var nextCursorMark = com.toolsverse.config.SystemConfig.instance().getProperties().
get("NEXT_MARKER");

value = !com.toolsverse.util.Utils.isNothing(nextCursorMark) ? nextCursorMark : null;

Available variables

The following variables can be referenced by name from JavaScript and Python code:

Programmatically set global and flow variables

Use loopKeyValues to programmatically set global and Flow variables as key=value pairs where both key and value are strings. Example: 

loopKeyValues.put('VAR_NAME', 'var_value');

Execute Flow fixed number of times

This variation of the Script loop executes a flow fixed number of times.

Step 1. Configure a Script loop.

Step 2. Enter the following Loop Script (example):

// this flow will be executed 40 times
value = evals <= 40 ? evals : null;

Database Loop . Read records from a database and execute the Flow for each record

The idea is that you can write a SELECT SQL statement which, when executed, creates a cursor driving the loop. The Flow in a database loop is executed as many times as the number of records returned by the cursor. For every column in each record in a cursor, the loop automatically sets the global and Flow variables.

Configure database loop

Step 1. Create a nested Flow and add steps (inner Flows).

Step 2. Select the Flow to be executed in a loop and go to Flow parameters.

Step 3. Select SQL in Loop Type.

Step 4. Select the database Connection used for the SQL loop.

Step 5. Enter the SELECT SQL statement to drive the loop in the Loop Script field. Optionally enable Execute as Script if the loop SQL is in fact script, such as PL/SQL, TSql, etc. Note that script must return a cursor. 

Example of the database loop

Let's assume that there is a role table in the database that contains information about roles and permissions. In this example, we are going to be creating CSV files for each role in a database.

Step 1. Create a database Connection that will be used to execute the driving (loop of) SQL.

Step 2. Create any Flow, for example, a Flow that reads data from a database and creates files, or a Flow that executes a web service. We're assuming that you want to create a loop that executes this Flow for each record in the result set.

Step 3. Create a nested Flow and add the Flow created in Step 2.

Step 4. Click the pen icon, next to the Flow name. The Parameters pop-up window will open.

Step 5. Select SQL for Loop Type.

Step 6. Select the Connection created in Step 1 for the Connection.

Step 7. Enter the driving SQL, for example, select roleid from role as the Loop Script .

Step 8. Modify the Flow created in Step 2 to use fields from the result set. Enclose the field name in curly brackets: {field name}. For example: {roleid}.

You can use the {field name} in:

• the FROM and TO fields for a transformation

• an HTTP Connection URL (or any other Connection property)
• anywhere in the source or destination query

When referencing {parameters}, make sure you are using the same CaSe for the field name as returned by the database. For example, some databases enforce UPPERCASE for column names returned by executing the SQL query: select roleid from role --> ROLEID, while other databases keep the original: select roleid from role --> roleid.

Debug database loop

Read how to debug a database loop.

File Loop. Read file names matching a wildcard and execute the Flow for each file

There is an easy way to ETL files matching wildcard.

Use file loop if want to execute more complex scenarios than just Extract-data-from-source-transform-and-load-in-destination.

Configure file loop

Step 1. Create a nested Flow and add steps (inner Flows).

Step 2. Select the Flow to be executed in a loop and go to Flow parameters.

Step 3. Set the Loop Type to Files by wildcard.

Step 4. Select the file Connection used for the loop.

Step 5. Enter the wildcard filename to drive the loop in the File path field.

File path modifiers

The File path parameter can include modifies.

• ;-file1,file2,filen: a comma separated list of files to exclude. Filenames in the exclude list can be wildcards. Example: *.csv;-test1.csv,test2.csv - process all files which match the wildcard *.csv, except files test1.csv and test2.csv.
• ;+file1,file2,filen: a comma separated list of files to include. Filenames in the include list can be wildcards. Example: *.csv;+test1.csv,test2.csv - process files test1.csv and test2.csv, ignore all other files which match the wildcard *.csv.
• [folder]: only include folders. Example *test- process folders which match the wildcard *test.
• ;oldest: sort files from oldest to newest. 
• ;newest: sort files from newest to oldest.
• ;ascdending: sort files alphabetically with ascending order.
• ;descending: sort files alphabetically with descending order.
• ;largest: sort files from largest to smallest.
• ;smallest: sort files from smallest to largest.

Sort modifiers override sort order set at the connection level.

Global and Flow variables set by the file loop

The following global and Flow variables are automatically set in each iteration of the file loop:

• loop_file_name: the name of the file processed by the loop. The name includes the extension but does not include the path.
• loop_base_file_name: the base name of the file processed by the loop. The name does not include the extension and the path.
• loop_file_id: the id of the file. Currently file id is only set by Google Drive connector. If there is no file id loop_file_id =loop_file_name. 
• evals: 1-based number of loop iteration.

Example of the file loop

Step 1. Create a Connection for the source files. Make sure the Connection supports wildcard filenames. Store the name of the Connection.

Step 2. Create a Flow where the source Connection is the Connection created in Step 1 and the destination is whatever you like.

Step 3. Create a nested Flow and add the Flow created in Step 2.

Step 4. Click the pen icon next to the Flow name and the Parameters pop-up window will open.

Step 5. Select Files by wildcard for Loop Type.

Step 6. Select the Connection created in Step 1 for Connection.

Step 7. Enter a wildcard filename, such as *.csv, for Files path.

Files will remain in the source folder after processing, so you might want to delete or move them to a different folder.

Parallel loop. Execute loop iterations in parallel threads

Loop iterations can be executed in parallel threads. For example, if the loop is configured to execute a flow which calls the HTTP endpoint with different parameters (e.g. start and end dates) you can configure the loop to make multiple HTTP calls in parallel, which can greatly improve the performance.  

Not all flow types are thread safe.

To enable parallel loop set the maximum number of the Loop Threads to use when executing the Flow in a loop. If the number is greater than one, each iteration of the loop will be executed in its own thread.

Retry loop iteration in case of error

When configuring a loop, you can set parameters for retrying a loop iteration in case of any error:

• Loop retries: number of retries for each loop iteration in case of error.
• Retries delay: number of milliseconds to wait between retries.

Stop the loop after the fixed number of iterations

You can configure any loop to stop after the fixed number of iterations by setting parameter Maximum number of iterations. The default is 100000.

Stop the loop programmatically 

Problem

The loop is running until there are records in the result set returned by the loop query, or until there are files to process, or until the JavaScript condition returns a not null value. In some cases, you may want to stop the loop early, based on conditions. 

Solution

Anywhere in the JavaScript or Python code, executed within the loop (for example in After Extract transformation) add the following line of code. This call is thread-safe.

etlConfig.setRequestStop(true);

Example

if (dataSet.getRecordCount() == 0) {
   etlConfig.log("Empty dataset, hence exiting the loop");

   etlConfig.setRequestStop(true);
}