mygemsSay hi

Merge CSV files into one

Several files in, one file out, columns matched by name.

  • Matches columns by name
  • Any number of files
  • Reports column mismatches
  • Nothing uploaded

1.Add the files

CSV or TSV. Several at once, or a few at a time.

Why it exists

Merging CSVs by hand goes wrong in one specific way: the files agree on what the columns mean but not on what order they are in, and a naive concatenation silently shifts every value one column left. This matches columns by name, tells you which files were missing which, and fills the gaps with blanks rather than with the wrong data.

What it does

  • Takes any number of CSV or TSV files and stacks their rows into one.
  • Aligns columns by header name, not by position, so column order can differ between files.
  • Names every column that is missing from at least one file, before you commit to the merge.
  • Optionally adds a `source_file` column so you can tell afterwards where a row came from.
  • Downloads the result as a single CSV.

What it will not silently do

It will not drop a column that only some files have, and it will not guess that two differently-named columns mean the same thing. Both of those are decisions, and it asks rather than assuming.

Questions

What happens if the files have different columns?
Every column from every file appears in the result. A row from a file that did not have a given column gets an empty value there. The mismatched columns are listed before you merge, so nothing about it is a surprise.
Does column order have to match?
No. Columns are matched by header name. This is the difference that matters — concatenating files by position when the order differs corrupts every row, quietly.
How many files can I merge?
As many as the tab has memory for. The practical limit is total size rather than file count.

The other tools