CSV To Dataset ID For Connector OS

Turn any CSV into an Apify Dataset ID — the universal input format for Connector OS Station.

If it's in a spreadsheet, it's in the pipeline. SalesNav exports, Apollo pulls, conference attendee lists, CRM dumps, spreadsheets from a VA — all become a dataset ID in seconds.

---

How to use it

This actor takes exactly one input. Fill in either the CSV Content field or the CSV URL field — not both. If both are filled, CSV Content wins and the URL is ignored.

Option 1 — Paste CSV text (CSV Content field)

Use this when you have the CSV in your clipboard.

• Open the CSV in a text editor, select all, copy, paste into the "CSV Content" textarea.
• The first row must be the column headers. Remaining rows are data.
• Works with any delimiter (comma, semicolon, tab, pipe) — auto-detected.
• Handles multi-line quoted cells, embedded commas, BOM.

Do not paste a URL into this field. The actor will detect it and fail with a clear error asking you to use the URL field instead.

Option 2 — Provide a link (CSV URL field)

Use this for CSVs hosted online or in Google Sheets.

• Direct .csv URL: paste it as-is.
• Google Sheets: paste the share link. The sheet must be shared as "Anyone with the link" (viewer access). If sharing is restricted, Google returns a login page and the actor fails with Got HTML instead of CSV.
• Only the first tab of a Google Sheet is exported unless the URL contains gid=<tab id>. Put your lead data on the first tab, or copy the URL from your browser while viewing the right tab (it contains #gid=...).

---

What comes out

Every row in the CSV becomes one item in the resulting Apify dataset. Columns are auto-detected from the header row. No filtering, no deduplication, no enrichment — pure pass-through.

• Dataset row count = CSV row count (minus the header row and empty lines).
• Plug the dataset ID straight into Connector OS Station.

---

Troubleshooting

The actor fails hard (does not create a half-full dataset) when anything at the input or output boundary is wrong. Every error message names the field and what to fix.

Symptom	Likely cause	Fix
Run fails: It looks like you pasted a URL into the "CSV Content" field	URL pasted into the wrong field	Put the URL in the "CSV URL" field, leave "CSV Content" empty
Run fails: The "CSV Content" field contains only whitespace	Textarea has only spaces, tabs, or blank lines	Paste real CSV text, or switch to the "CSV URL" field
Run fails: The "CSV URL" field must be a full http(s) URL	Missing scheme (e.g., docs.google.com/... instead of https://docs.google.com/...)	Include https:// at the start of the URL
Run fails: Got HTML instead of CSV. If Google Sheets, make sure the sheet is shared...	Google Sheet is restricted or requires login	Share → General access → "Anyone with the link" (Viewer)
Run fails: Failed to fetch CSV: HTTP 4xx/5xx	URL is wrong, private, or the host is down	Open the URL in a private browser window; if you cannot access it anonymously, neither can the actor
Run fails: Fetched URL returned an empty body	Link resolves but the server sent nothing	Try the link in a browser; confirm it downloads a non-empty file
Run fails: CSV parsed but contains no data rows	Content has a header row only (no data), or the file is empty after parsing	Paste the full CSV including data rows; for Google Sheets, confirm the exported tab has data on it
Log warning: only 1 column detected	papaparse auto-detection misread the delimiter (or the file is genuinely single-column)	If multi-column was expected, check the source delimiter; for TSV/pipe-delimited files, re-export as comma-delimited
Log warning: Parse warnings: N	CSV has unescaped quotes or inconsistent column counts — some rows may have collapsed	Open the CSV in a text editor, look near the error row number printed, fix the quoting
Fewer rows than expected (but no warning)	CSV contains multi-line quoted cells — one row legitimately spans many raw lines	Count rows from the source (spreadsheet row count), not from the raw file line count

---

What the log tells you

On every run, look at the first few log lines:

• CSV from direct input (N chars) — you used the CSV Content field.
• Fetching CSV from <url> — you used the CSV URL field.
• Google Sheets detected → CSV export URL — the URL was rewritten to the Sheets export endpoint.
• Google Sheets URL has no gid — defaulting to first tab (gid=0)... — if your data is on a different tab, include #gid=<id> in the URL (the browser address bar shows the gid when you click the tab).
• N rows, M columns: col1, col2, ... — what was parsed. This N is the dataset item count.
• Parse warnings: N (if any) — the CSV has quoting issues and some rows may have collapsed. Inspect the source CSV.
• WARNING: only 1 column detected — delimiter detection likely failed. Check the source file.

If the run fails, the error message tells you exactly what went wrong — this actor does not silently produce empty or partial datasets.

---

Tech notes

• Runtime: Node 20.
• Parser: papaparse — battle-tested against millions of CSVs.
• Google Sheets URL rewrite: docs.google.com/spreadsheets/d/<ID>/... → export?format=csv&gid=<gid or 0>.
• Output: one Actor.pushData call with the full row array.