Creating Flows with Loops

1. Create a flow and select the step that you want to use as a loop step.
2. Click the Create Loop button in the authoring pane toolbar.

3. From the menu that appears, select Set as loop.

   

4. For a step that has been set as a loop, the Step Properties > Inputs tab has an extra Loop Details section, where you can define the collection that the input will be iterated on and the break points of the loop.

   

   • For: Enter the name of the variable that will be assigned the current value from the collection that is defined under In. This step is compulsory.

     This variable will be available while the loop step is running.

   • In: Define the collection to be iterated on. This step is compulsory.

     You can define the collection using a flow variable or system property. For example, you can create a system property with a comma delimited string ('1,2,3,4,5') as the value and use this to define the collection.

     Note The evaluation of the flow variable or system properties must be a string as CloudSlang supports only strings. For example. 'Range(1,6)' will work inline only if the for variable (which will be a digit) isn't assigned to an inner step input (which requires string). It will not work if passed as a value of system property or flow variable.

     You can also open the Expression Editor and write a Python expression that results in a list or map. For example:

     • An array: ['1','2','3','4'] or ['a','b','c','d']

     • A Python expression: map(str,range(1,5))

     • A map: {"key1":"value1","key2":"value2"}

     • A comma delimited string, which will be split automatically into a list: '1,2,3'.

   • Break On: A loop ends either when the list that it is looping on is finished or when the looped step returns one of the results defined as a break point.

     Click the Select break result to add button to choose one of the step results as a break point. There can be multiple break points.

     Note To remove a break point, roll over it so that the Remove button is displayed, and click the Remove button.

     

5. In the Step Properties > Outputs tab, define how the output will be published. At run time, the output will contain the value from the last iteration of the loop.

   1. Click the Add a Step Output button to add a new output to the step.
   2. Choose one of the underline outputs to define the output.

   3. Under Value, open the Expression Editor and enter a Python expression to aggregate all the iterations.

   4. Set up the navigation logic for the loop step.

      The navigation to the next step in the flow occurs when the last iteration of the step is completed or after the loop exits the iteration due to a break. The navigation will run according to the last iteration of the loop.

      Note If you want to use the for value inside the step, you must add it as the value to one of the step inputs.

To aggregate the results of a specific loop step, proceed as in the following example, where we add the numbers generated in the loop's iterations:

1. Import a base content pack onto dependencies pane and choose the random_number_generator operation.

2. Drag the random_number_generator operation to the flow and set the resulting step as a loop.

3. Define the collection that the loop will run on as shown in the following example image.
   

   

4. Enter the values for min and max Step Inputs as shown in the following example image.

   

5. Define the aggregated output for the loop step as shown below.
   

   Note The random_number output shown in the above image is optional.

   
   The Python expression for the sum_random output using (aggregating) the outputs from the iterations is:
   
   
6. Add a new Step Input and rename it as the Output. For example the following image shows the Step Input name as sum_random. This sum_random step input is assigned from itself, or with 0 when it has no value yet (at the first iteration).
   
7. Click Save . The flow is saved.

In a parallel loop step, a new branch is created for each value in the collection and the branches run in parallel. The rest of the flow will continue only after all the branches have finished running.

1. Create a flow and select the step that you want to use as a parallel loop step.
2. Click the Create Loop button in the Authoring pane toolbar.

3. From the menu that appears, select Set as a parallel loop.

   

4. For a step that has been set as a loop, the Step Properties > Inputs tab has an extra Loop Details section, where you can define the collection that the step will iterate on.

   

   • For: Enter the name of the variable that will be assigned the current value from the collection that is defined under In. This step is compulsory.

     This variable will be available while the loop step is running.

   • In: Define the collection to be iterated on. This step is compulsory.

     You can define the collection using a flow variable or system property. For example, you can create a system property with a comma delimited string (1,2,3,4,5) as the value and use this to define the collection.

     You can also open the Expression Editor and write a Python expression that results in a list or map. For example:

     • An array: ['1','2','3','4'] or ['a','b','c','d']

     • A Python expression: map(str,range(1,5))

     • A map: {"key1":"value1","key2":"value2"}

     • A comma delimited string, which will be split automatically into a list: '1,2,3'.

5. In the Step Properties > Outputs tab, define how the output will be published.

   1. Click the Add a step output button to add a new output to the step.

   2. Choose one of the underline outputs to define the output:

      

   3. Under Value for the output, open the Expression Editor and enter a Python expression to aggregate the outputs.

      

      To define the step aggregation logic, you can use the branches_context construct. This is a list that includes the context of all the iterations. The branches_context is set according to the order of the executions, so that the execution that finishes first takes the first place in the list, and so on.

      After all the branches of the parallel loop have completed, the execution of the flow continues with the output that is published. The expression of each output pair is evaluated and published to the flow's scope. The out-of-the-box outputs are name (first branch), name_list, iteration_size, result_status_list, and num_of_failures.

      Examples:

      • To retrieve the name variable from the first branch to finish:

        branches_context[0]['name']

      • To publish aggregated output: (this iterates over the branches_context, which contains a list of all the iterations of the parallel flow, and takes the value of 'name' from each branch)

        str([str(x["name"]) for x in branches_context])

      • To create a list of branch results: (as branches of a parallel flow complete, the branch results get placed into the branches_context list under the branch_result key)

        str([str(x["branch_result"]) for x in branches_context])

      • To configure the iteration size.

        str(len(branches_context))

      • To configure the number of failures, use num_of_failures:

        str([x["branch_result"] for x in branches_context].count("FAILURE"))

1. Select the loop step.
2. Click the Create Loop button in the Authoring pane toolbar.

3. From the menu that appears, click the selected option to clear it.