What to Do When a CSV Is Too Large for Google Sheets

If your file is under the cell limit and still fails to import, the 100 MB cap may be what stopped it, so use the splitter’s maximum-size mode instead. You can also cut the file down to the rows and columns you need rather than split it. For data far beyond the limit, use Connected Sheets with BigQuery, or a tool built for large files.

You import a CSV into Google Sheets and it fails, or Google tells you “This file is too large to import directly into Google Sheets”. The file may be nowhere near as big as ones Excel opens, because Sheets does not count rows the way Excel does. It counts cells.

The Limit Is Cells, Across the Whole Spreadsheet

Google’s help centre gave the limit as “up to 20 million cells or 100MB for spreadsheets that are created in or converted to Google Sheets” when we read it on 10 October 2026, with no separate column limit. Cells means rows multiplied by columns, and the allowance is for the whole spreadsheet, so every tab draws on it.

That limit is 20 million cells now. It was 10 million before the increase Google announced on 10 September 2026, which was scheduled to reach the last Workspace business domains by mid-October, so older answers and screenshots still show 10 million. Google says the new limit “applies to new, existing, and imported files”, for Google Workspace business customers, Workspace Individual subscribers and personal Google accounts.

The 100 MB cap applies to the spreadsheet created in or converted to Sheets rather than to the CSV, so it does not translate into an exact file size to stay under, but a file with long values can fail before it reaches 20 million cells. The same announcement said Google plans to “increase the file byte size limit for imported spreadsheets”, without giving a date or a new figure.

Will Your File Fit?

Multiply the number of rows, including the header, by the number of columns. If the answer is under the limit, the cells fit. Some examples:

The column count is the number of names in the header line. For the row count, use a tool that parses the CSV, not a line counter: a value containing a line break spans two lines but is one row. Counting the rows in a CSV shows how, and RowSlice’s CSV Analyzer reports the data row count and the column list in one pass, which is both halves of this multiplication without opening the file anywhere. A separate reference sets this cell limit beside the maximum rows and columns of other programs and importers, each with its source. To see what Sheets does near its limits before a real file hits them, generate a sample CSV file of the size you need.

Import into a new spreadsheet rather than a new tab of an existing one, so no other tab uses up the allowance. In File › Import, choose Create new spreadsheet as the import location. The same dialog has a Separator character setting — Detect automatically, Tab, Comma or Custom. Google’s import help page says automatic detection can recognise semicolons, but if the data lands in a single column, import again and choose the separator yourself.

Option 1: Split It into Parts That Fit

The simplest fix is several spreadsheets instead of one. Work out the rows per part from the table above, leave a margin, and split. A 20-column file fits up to 1,000,000 rows, so about 900,000 rows per part leaves headroom. Parts that size are also under Excel’s row limit of 1,048,576, so they open whole in Excel as well.

RowSlice can split the CSV into multiple files in your browser: choose Rows per file, enter the number, and every part gets the header row. It parses the file properly, so a line break inside a value never cuts a row in two, and the file is not uploaded anywhere. If the cells fit and the import still failed, the 100 MB cap may be what stopped it, so use the maximum-size mode instead. Then import each part with Create new spreadsheet.

Option 2: Import Only the Part You Need

Often nobody needs every row and column in Sheets. Four ways to cut the file down before importing:

• Only the rows that meet a condition. RowSlice’s CSV filter keeps the rows where a column equals or contains a value, is empty, or falls in a number range — one status, one country, orders over a given amount — and writes those rows to a new file with the header on top.
• One file per team, region or client. Splitting the CSV by column value writes one file for each value in a column, each far smaller than the whole.
• Fewer columns. Every column you drop multiplies into many fewer cells: removing 10 of 20 columns doubles the rows that fit. Reducing a CSV’s size has tested commands for keeping only the columns you name.
• No repeated rows. Exports that were merged or re-run often contain duplicates, so remove duplicates from the CSV before you import it.

Option 3: Connected Sheets and BigQuery

Google’s own answer for data beyond the limit is Connected Sheets, which keeps the data in BigQuery, Google’s data warehouse, and lets you work with it from a sheet. Google’s Connected Sheets help says you can “access, analyze, visualize, and share billions of rows of data” this way. Since November 2025, opening a large CSV from Drive gives “a new option to import with Connected Sheets when a file is too large for native upload to Sheets”, according to Google’s announcement.

It is not the same as having the rows in a sheet. You analyse them through pivot tables, charts and formulas, and Google caps what comes back: “up to 200k rows for pivot tables” and “up to 500k rows or 5m cells for extracts”. The requirements include access to Google Cloud and BigQuery and “a project with billing setup in BigQuery”, and an administrator may need to allow it. It suits a team that already uses BigQuery, and is a lot of setup for one file.

Option 4: Use Something Built for Large Files

If the goal is to look through the data, filter it or total it, a spreadsheet may be the wrong tool at this size. When looking is all you need, RowSlice’s CSV Viewer pages through the rows and searches them without importing anything anywhere, which costs nothing against a cell limit because there is no spreadsheet. DuckDB answers questions about a multi-gigabyte CSV with one line of SQL, and Excel’s Data Model holds up to 1,999,999,997 rows per table. Opening a CSV that is too large for Excel compares these and other options; most of them apply equally when the spreadsheet you had in mind was Google Sheets.