Hi Zapier Community,

I'm working on a Zap designed to automatically process a CSV file received daily as an email attachment and use it to update or create records in a Zapier Table. I've gotten close, but I'm running into some confusing issues, particularly with testing steps inside a loop.

My Goal:
Receive daily email -> Parse CSV attachment -> For each row: Find record in Zapier Tables by entityInterface, if found Update, if not found Create.

My Current Workflow:

1. Trigger: Email Parser by Zapier (New Email - gets attachment)

2. Action: Code by Zapier (Run Python) - Fetches the attachment content via URL and returns the entire CSV content as a single text string (e.g., in an output field named csv_content). (Reason: Tried outputting a list/line items directly, but hit the 250 item limit).

3. Action: Looping by Zapier - Create Loop From Text

   • Input Text: Mapped to csv_content from Step 2.

   • Text Delimiter: Set to \n (newline).

4. Action (Inside Loop): Code by Zapier (Run Python)

   • Input full_text: Mapped to csv_content from Step 2.

   • Input line_number: Mapped to Loop Iteration from Step 3.

   • Code: Parses the specific line (linessline_number - 1]) from full_text using csv.reader and outputs a dictionary with keys like entityName, entityInterface, Description, OperationalStatus.

5. Action (Inside Loop): Zapier Tables (Find Record)

   • Table: My Target Table

   • Lookup Field: Field 2 (This column holds my unique entityInterface values)

   • Lookup Value: Mapped to 4. Run Python -> entityInterface.

6. Action (Inside Loop): Paths

   • Path A (Found): Rule checks if 5. Find Record -> Zap Search Was Found Status is true.

   • Path B (Not Found): Rule checks if 5. Find Record -> Zap Search Was Found Status is false.

7. Action (Inside Path A): Zapier Tables (Update Record)

   • Record ID: Mapped to 5. Find Record -> Record ID.

   • Field 1 (entityName): Mapped to 4. Run Python -> entityName.

   • Field 3 (Description): Mapped to 4. Run Python -> Description.

   • (Field 2 / entityInterface is NOT mapped here)

8. Action (Inside Path B): Zapier Tables (Create Record)

   • Field 1 (entityName): Mapped to 4. Run Python -> entityName.

   • Field 2 (entityInterface): Mapped to 4. Run Python -> entityInterface.

   • Field 3 (Description): Mapped to 4. Run Python -> Description.

The Problem / Confusion:

When I test the steps inside the loop (Steps 4, 5, 7, , Zapier seems to consistently use data from the first iteration (Loop Iteration = 1).

• This causes the Step 4 (Run Python) test to show it's outputting the header names (entityName, entityInterface, etc.) as the values.

• This then causes the Step 5 (Find Record) test to fail ("Nothing could be found") because it's searching Field 2 for the literal text entityInterface.

• This then causes the Step 7 (Update Record) test to fail ("Cannot create ULID...") because it receives an invalid Record ID from the failed Step 5 test.

My Questions:

1. Is the current workflow approach (Code -> Loop From Text -> Code -> Find -> Paths -> Update/Create) a valid and robust way to handle this, given the Code step item limit?

2. Is it normal for the tests of steps inside a loop to consistently use Iteration 1 data, even if prerequisite steps failed their tests based on that data?

3. Are there any obvious flaws or better ways to configure this?

I feel like the logic is almost right, but the testing experience inside the loop is making it hard to be 100% confident. Any insights or suggestions would be greatly appreciated!

Thanks!