Data cleaning for CRM imports

HubSpot import errors

Fix "incorrect number of columns" on a HubSpot import

This error means HubSpot counted a different number of fields in a row than in the header row. A CSV marks each column with a comma, so almost always this is a quoting problem: a value contains a comma or a line break that wasn't wrapped in double quotes, and the row splits into the wrong number of fields. Fix the structure and the import runs clean.

Fix your CSV in Sift, free →

What the error looks like

This is the highest-traffic mismatch in HubSpot's import community: a row that looks fine in a spreadsheet is ragged in the raw CSV underneath.

Why it happens

• Unquoted commas inside a field. Company names, addresses, and job titles often contain commas. Without double quotes around the value, each comma reads as a new column.
• Line breaks inside a cell. A multi-line note or address carries a newline that ends the CSV row early and splits it in two.
• Stray trailing delimiters. A comma left at the end of a row adds a phantom empty column that the header row doesn't have.
• Mixed delimiters. Part of the file uses commas and part uses semicolons or tabs, so the field count drifts row to row.
• A merged or blank cell shifting columns. A cell that was merged in the source, or a missing separator, pushes later values into the wrong columns.

The manual fix

1. Open the raw CSV in a plain-text editor, not just a spreadsheet, so you can see the actual commas and quotes.
2. Find the rows whose field count differs from the header row; those are the ragged ones.
3. Wrap every comma-containing field in double quotes, so Acme, Inc becomes "Acme, Inc" and its comma stays inside one column.
4. Remove in-cell line breaks, or wrap the whole multi-line value in double quotes so the break is kept inside the field.
5. Re-save the file as UTF-8 CSV and try the import again.

The Sift fix

1. Drop your CSV or Excel file into Sift. It runs in your browser; the file is never uploaded.
2. Sift re-parses the CSV robustly and tidies the column headers, so the header row and the data rows line up.
3. It re-exports a clean, single-sheet CSV with correct quoting, so every row has exactly the header's column count and the "incorrect number of columns" error goes away.
4. Any row that is genuinely ragged, where the intended value is ambiguous, is flagged to a "needs your eyes" list for you to review instead of being silently guessed.

What Sift can't fix

Honesty matters here: Sift fixes structure, not missing intent. If a row is genuinely missing a value, or two columns were merged in the source so the data is truly ambiguous, Sift re-parses what's there and flags the row, but a human still has to decide what the intended value was. It can line the columns back up; it can't invent the data that should have been in them.

Related guides