paste

What it does

paste merges corresponding lines from two or more files horizontally: line 1 of each file becomes one output line (joined by tabs by default), line 2 of each file becomes the next output line, and so on. It's essentially the transpose of cat, which stacks files vertically — paste lines them up side by side as columns, which is useful for quickly assembling simple tabular data from separate single-column files.

Beginner examples

• paste names.txt scores.txt — combine two files into two tab-separated columns
• paste -d, file1.txt file2.txt — use a comma instead of a tab as the delimiter
• paste file1.txt file2.txt file3.txt — merge three files into three columns
• paste -s file.txt — join every line of a single file into one line, tab-separated

paste usernames.txt emails.txt

Advanced examples

• Build a quick CSV from two single-column files: paste -d, names.txt scores.txt > people.csv
• Flatten a column of values into a single comma-separated row: paste -s -d, ids.txt
• Combine paste with process substitution to merge command output without temp files: paste <(cut -d, -f1 data.csv) <(cut -d, -f3 data.csv)
• Use multiple delimiters that cycle across columns: paste -d ':,' file1.txt file2.txt file3.txt (colon between first pair, comma between next, then repeats)
• Merge sorted key-value files where positions align exactly (when a proper join isn't needed): paste keys.txt values.txt

paste <(cut -d, -f1 data.csv) <(cut -d, -f3 data.csv) > subset.csv

Common mistakes

• Using paste when the files aren't line-aligned — since it merges purely by line position, mismatched files produce nonsensical rows with no warning.
• Reaching for paste when you actually need to merge on a matching key field, which is join's job, not paste's — paste has no concept of matching values, only line position.
• Forgetting the default delimiter is a tab, which can look like several spaces in a terminal and cause confusion when the output is later parsed as comma-separated.
• Using -s when you meant the default column-merging mode (or vice versa) — they produce very different shaped output.

Tips

• Use process substitution (<(...)) to paste together columns extracted on the fly from a single source file, without creating temporary files.
• paste -s -d, file is a quick way to turn a newline-separated list into a single comma-separated line, handy for building shell arguments or simple reports.
• Check that all input files have the same number of lines before pasting, since a shorter file simply leaves empty fields for its missing lines rather than raising an error.

Best practices

• Use paste for simple, positionally-aligned column assembly; use join instead as soon as you need to match rows by a key rather than by line number, since mismatched ordering with paste silently produces wrong data.
• Sort and verify line counts of input files before pasting when building anything more than a throwaway one-off, to avoid silent misalignment.
• Prefer -d with an explicit delimiter over relying on the default tab when the output will be consumed by another tool expecting a specific format like CSV.

Real-world use cases

• Quickly building a two-column report from separately generated lists, such as usernames and their corresponding email addresses.
• Turning a column of IDs into a single comma-separated string to embed in a SQL IN (...) clause: paste -s -d, ids.txt.
• Reassembling extracted columns from a larger CSV after processing them independently with cut.

Common interview questions

• What does the paste command do? It merges corresponding lines from two or more files side by side into columns, joined by a delimiter (tab by default).
• What's the difference between paste and join? paste merges purely by line position with no regard for content; join merges rows based on matching a common key field, similar to a database join, and requires sorted input on that key.
• How would you turn a file's lines into a single comma-separated line? paste -s -d, file.txt, which uses serial mode to merge all lines of the file into one, separated by commas.