fix(trigger): fix Google Sheets trigger header detection and row index tracking

[waleedlatif1]

Summary

• Auto-detect header row by scanning the first 10 rows for the first non-empty row, instead of hardcoding row 1 — fixes `row: null, headers: []` for sheets where headers aren't in row 1
• Replace `lastKnownRowCount` with `lastIndexChecked` (1-indexed row pointer) for accurate new-row detection
• Skip header row from data emission using `adjustedStartRow` so it never fires a workflow run
• Use full A:Z fetch for sheet state so row count is accurate regardless of which columns have data
• Preserve `lastModifiedTime` when remaining rows exist beyond the batch cap, preventing the Drive pre-check from skipping them on the next cycle
• Skip empty rows in `processRows` to avoid firing workflow runs with no data

Type of Change

• Bug fix

Testing

Tested manually

Checklist

• Code follows project style guidelines
• Self-reviewed my changes
• Tests added/updated and passing
• No new warnings introduced
• I confirm that I have read and agree to the terms outlined in the Contributor License Agreement (CLA)

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

1 Skipped Deployment

Project	Deployment	Actions	Updated (UTC)
docs	Skipped		Apr 11, 2026 6:30pm

[cursor]

PR Summary

Medium Risk
Changes polling state semantics (lastKnownRowCount -> lastIndexChecked) and header/row filtering, which could cause one-time missed or duplicate events for existing webhooks if stored provider config isn’t migrated or edge cases exist around blank/header rows.

Overview
Fixes Google Sheets polling to auto-detect headers (scan first 10 rows for the first non-empty row) and fetch sheet state from A:Z so row counting doesn’t miss data in non-A columns.

Replaces lastKnownRowCount with a 1-indexed lastIndexChecked pointer, ensures the detected header row is never emitted as a data event, skips empty rows during processing, and preserves lastModifiedTime when a poll hits the per-batch cap so remaining rows aren’t skipped next cycle.

^{Reviewed by Cursor Bugbot for commit 9b3af87. Configure here.}

[greptile-apps]

Greptile Summary

This PR fixes several correctness issues in the Google Sheets polling trigger: auto-detecting the header row (scanning up to 10 rows instead of hardcoding row 1), replacing the old lastKnownRowCount counter with a 1-indexed lastIndexChecked pointer, skipping the header and pre-header blank rows from data emission via adjustedStartRow, preserving lastModifiedTime so the Drive pre-check doesn't skip remaining batches, and removing the spurious processedCount increment for skipped empty rows. All three concerns raised in the previous review threads have been addressed.

Confidence Score: 5/5

Safe to merge; all previously raised concerns are resolved and the updated logic handles edge cases correctly.

Thorough review of the single changed file found no P0 or P1 issues. The adjustedStartRow offset, pointer advancement by rowsToFetch, empty-row skip, lastModifiedTime preservation, and first-poll seeding all behave correctly across the edge cases examined (empty leading rows, header mid-sheet, partial-failure retry, batch cap with remaining rows). All three threads from the previous review cycle are resolved.

No files require special attention.

Important Files Changed

Filename	Overview
apps/sim/lib/webhooks/polling/google-sheets.ts	Rewrites header detection to auto-scan first 10 rows, replaces lastKnownRowCount with lastIndexChecked, skips header row from emission via adjustedStartRow, preserves lastModifiedTime for remaining batches, and drops empty-row processedCount increment; logic is correct across the edge cases reviewed.

Flowchart

%%{init: {'theme': 'neutral'}}%%
flowchart TD
    A([pollWebhook]) --> B[resolveOAuthCredential]
    B --> C[isDriveFileUnchanged]
    C -->|unchanged| D[updateConfig lastCheckedTimestamp\nmarkWebhookSuccess]
    C -->|changed| E[fetchSheetState A:Z\nrowCount · headers · headerRowIndex]
    E --> F{lastIndexChecked\nundefined?}
    F -->|yes — first poll| G[seed lastIndexChecked = rowCount\nmarkWebhookSuccess]
    F -->|no| H{currentRowCount\n<= lastIndexChecked?}
    H -->|yes — no new rows| I[update lastIndexChecked = currentRowCount\nmarkWebhookSuccess]
    H -->|no — new rows found| J[compute startRow · endRow · rowsToFetch\nadjustedStartRow = max startRow headerRowIndex+1]
    J --> K{adjustedStartRow\n> endRow?}
    K -->|yes — batch is all header/blank| L[advance pointer by rowsToFetch\npreserve lastModifiedTime if remaining\nmarkWebhookSuccess]
    K -->|no| M[fetchRowRange adjustedStartRow→endRow]
    M --> N[processRows\nskip empty rows\nfire workflow per row via idempotency]
    N --> O{failedCount > 0?}
    O -->|all failed| P[markWebhookFailed\ndo NOT advance pointer\nreturn failure]
    O -->|partial or none| Q[advance pointer by rowsToFetch\npreserve lastModifiedTime if remaining or failed\nmarkWebhookSuccess]

Loading

_{Reviews (3): Last reviewed commit: "fix(trigger): don't count skipped empty ..." | Re-trigger Greptile}

[waleedlatif1]

@greptile

[waleedlatif1]

@cursor review

[waleedlatif1]

@greptile

[waleedlatif1]

@cursor review

[cursor]

✅ Bugbot reviewed your changes and found no new issues!

Comment @cursor review or bugbot run to trigger another review on this PR

^{Reviewed by Cursor Bugbot for commit 9b3af87. Configure here.}