CSV Join Tool

Join two CSV tables by exact key values using inner, left or full joins with explicit duplicate-key handling.

Processing happens in this browser. Your CSV inputs are not uploaded.

Each input: up to 200,000 characters, 200 unique nonempty headers and 10,000 data rows. Maximum output: two million characters.

Keys match exactly and case-sensitively. Empty or whitespace-only keys never match. Expansion can multiply rows; output is capped at 10,000 records.

Headers receive left. and right. prefixes; both key columns are retained. Blank unmatched cells are exported as empty strings.

Preserve mode keeps cell strings. Protection prefixes cells beginning with =, +, - or @ after whitespace with an apostrophe, including negative numbers. This changes values. Quoting and line endings are normalized.

Join records from two CSV exports

CSV Join Tool combines columns from two tables by comparing one key column from each. Paste the left and right CSV, choose the corresponding key columns, select a join mode and choose Join tables. The result includes every column from both inputs, including both keys. Headers receive left. and right. prefixes so their origin stays visible even when the original names overlap.

Example: attach a department

Load the examples. The left table contains Ada with ID 1 and Lin with ID 2. The right table contains ID 2 in Sales and ID 3 in Support. With the default Inner: matched records only mode, only Lin's row joins, with Sales appended. Left mode also keeps Ada, with blank right-hand cells. Full mode additionally appends ID 3 from the right table with blank left-hand cells.

Matching and duplicate keys

Matching is exact and case-sensitive: 01 differs from 1, and spaces around a nonblank key are significant. Empty or whitespace-only keys never match, including other empty keys. They can appear as unmatched records in left or full mode.

The default duplicate policy rejects repeated nonempty keys in either table rather than silently choosing one record. Expand every matching pair allows one-to-many and many-to-many output. If a key occurs twice on the left and three times on the right, it creates six joined rows. All left rows remain in their original order, and matches follow right-table order. Full mode appends unmatched right rows in their original order. Blank unmatched cells cannot be distinguished from actual empty values by the CSV alone; keep key columns when interpreting them.

Boundaries and errors

Each input supports 200,000 characters, 200 unique nonempty headers and 10,000 data rows. Output is limited to 10,000 records and two million characters. Large duplicate expansions stop with an error, not a partial download. There is no fuzzy matching, type conversion, composite key, database query or automatic aggregation.

The output includes both sets of columns, so joining two 200-column inputs can produce 400 columns. Other header-aware tools here accept at most 200 columns. If you plan to use one afterward, select only the necessary columns in each input before joining so their combined width stays within that limit.

Preserve export keeps cell strings and normalizes CSV quoting and line endings. Spreadsheet protection instead prefixes formula-like values with an apostrophe, including negative numbers. All processing happens locally; no table is uploaded. Editing either input or an option clears stale output.

Use CSV Row Filter before joining to reduce records, and CSV Column Selector afterward to remove redundant key columns. Use CSV Merger when you want to append rows with matching schemas rather than combine columns by key.