Combine two or more columns in a CSV

How to combine columns in a CSV

1. 1. Drop your CSV into the tool. The file is read into your browser’s memory — it is never sent over the network.
2. 2. Pick the columns to combine. Tick two or more columns, such as first_name and last_name. They are joined in the order you click them.
3. 3. Choose a separator and name the column. Use a space, comma, dash, underscore, or nothing at all, then give the new combined column a name like full_name.
4. 4. Preview and download. Check the new column in the preview, then create a free account to export the full CSV.

Why combine columns?

CRMs and email tools usually want a single field where your export has several — a full_name built from first and last name, or a single address line built from street, city, and postcode. Combining the columns before import means cleaner records and fewer mapping errors. Because Sigmera processes the file locally with the browser’s built-in JavaScript engine, there is no server to upload to and nothing is stored — which satisfies GDPR data-minimization (Article 5) because zero bytes leave the device.

Combine CSV columns vs merge CSV files

“Combine CSV files” and “merge CSV files” usually mean stacking whole files — twelve monthly exports with the same headers becoming one long table. That is a row operation. This tool does the column operation: two or more fields inside one file joined side by side into a new field. Both end in a file that follows the same RFC 4180 record structure — one header line, then records with a matching field count — which is why the two jobs get confused.

If you need to stack files, do that in Excel or Google Sheets first (the merge-spreadsheets comparison rates the options), then bring the combined file here. A stacked file almost always contains the same person twice, so removing duplicate rows is the sensible next pass — and both passes run in the browser, so the merged list never gets uploaded.

Facing the reverse problem — one column holding whole names that the importer wants as two fields? The name splitter does that pass, including “Last, First” order and an optional middle-name column. Inside Excel, the equivalent is splitting first and last name with Text to Columns or a formula.

Where the TEXTJOIN formula stops working

Excel can do this with a formula, and for a one-off on a small sheet that is often the right call. It is worth knowing the three places the formula route breaks before you build a process around it.

• It is not in every Excel. Microsoft lists TEXTJOIN as available in Excel 2019, 2021, 2024 and Microsoft 365 — and it requires Office 2019 or a current subscription. Send a workbook built on it to someone running Excel 2016 and the cells come back #NAME?.
• Empty cells are an argument, not a default. TEXTJOIN’s second parameter is ignore_empty: set to TRUE it skips blanks, set to FALSE it writes the delimiter anyway. That single boolean is what separates Ada Lovelace from Ada  Lovelace on the rows with no middle name.
• There is a hard character ceiling. Microsoft documents that if the resulting string exceeds 32,767 characters — the cell limit — TEXTJOIN returns the #VALUE! error rather than a truncated join. Long concatenated address or note fields hit this.

The full syntax and version matrix are in Microsoft’s TEXTJOIN reference. None of the three applies here: this tool skips empty cells without an argument to set, has no cell-length ceiling because the output is a CSV field rather than a spreadsheet cell, and needs no particular version of anything — it runs in the browser tab you already have open, and the formula walkthrough covers the spreadsheet route in full if you would rather stay there.

The Google Sheets route, and where it differs from Excel

Sheets has the same three options — the & operator, CONCATENATE, and TEXTJOIN — but without Excel’s version problem. Google documents the signature as TEXTJOIN(delimiter, ignore_empty, text1, [text2, ...]), with ignore_empty as a boolean that drops blank cells from the result when it is TRUE. Every Sheets file has it, so the #NAME? that appears when an Excel 2016 colleague opens your workbook has no equivalent here.

Two things still bite. The formula produces a column of formulas, not values — export that sheet to CSV and you get the results, but copy it to another tab and the references break, so paste-special as values first. And the import is where a phone or postcode column loses its shape: Sheets applies its own type detection to a pasted or uploaded CSV, so the joined field can arrive already altered. Getting the file in without that happening is the subject of importing a CSV into Google Sheets, and the same caution applies to a dedupe pass run in the sheet. Joining the columns before the file reaches a spreadsheet skips both problems: there is no formula layer to flatten and no import step to survive.

Combining columns without Excel rewriting the values

The usual route to merge CSV files or join their columns is to open the file in a spreadsheet first, and that is where the damage happens. Microsoft states that Excel removes leading zeros and converts large numbers to scientific notation so that formulas and maths operations can run on them. A postcode of 01234 becomes 1234, and a sixteen-digit account reference becomes 1.23E+15 — before you have combined anything. Build a combined address or reference field out of columns that have already been rewritten, and the new field inherits every one of those errors.

This tool never parses the values as numbers. The CSV is read as text, the selected fields are concatenated as text, and the result is written back as text, with the RFC 4180 quoting applied only where a value actually needs it. If the file does have to reach a workbook afterwards, do the join here first and convert the finished CSV to .xlsx — the types are already settled in the file, so there is nothing left for Excel to infer.

A worked example: what happens to each cell

The combiner applies the same three rules to every row, so the result is predictable before you preview it. Here is a four-row export joined on first_name, middle, and last_name with a space as the separator.

• Each value is trimmed first. Leading and trailing spaces inside a source cell are removed before the join, which is why the padded “ Alan ” comes out as a clean Alan Turing. Spaces in the middle of a value are left alone.
• Empty cells are skipped, not joined. A blank middle name adds nothing — no double space, no stray comma — and a row where every selected cell is blank gets an empty combined value rather than a lone separator.
• The separator goes only between values that exist. That is the behaviour you get in a formula only by setting TEXTJOIN’s ignore_empty to TRUE; here it is the only behaviour.

The source columns are untouched and the new field lands at the right-hand end of the file. If the trimmed names now contain the same person twice — Grace Hopper from two different exports — run a duplicate check on the combined column next, so the match is made on the whole name rather than on two fields separately.

Joining columns is usually one step of a longer clean-up. The full set of online CSV tools covers the passes on either side of it — trimming, format standardising, deduplicating, and converting — and lists the order to run them in so a later pass does not undo an earlier one.