Skip to content

hash_columns

Append one column holding a hash of chosen columns of a tabular file and write the result. Same engine as Columns > Hash columns... in the app and octa --hash-columns.

This is a write tool, so it is removed under --mcp-read-only.

When to use

  • Building a hash key for a data warehouse load.
  • Giving rows a stable fingerprint to compare two extracts.

Input schema

Parameter Type Required? Default Description
path string yes (no default) Path to the source file
columns string[] yes (no default) Column names, joined in this order
algo string no md5 md5, sha256 or sha512
delimiter string no \| Put between the values; may be empty
null_text string no "" Stands in for a NULL cell
trim bool no false Strip whitespace from each value first
upper bool no false Upper-case each value first
new_column string no hash_<columns> Name of the new column; must not exist yet
output_path string no overwrite path Where to write the result; format follows its extension
unlimited bool no false Lift the file-loader row cap so every row is read and hashed

Every value is turned into text first, whatever its type. The hash is lowercase hex. A file read only in part (the row cap) is refused rather than written back with fewer rows; pass unlimited: true. Database files (SQLite / DuckDB / GeoPackage) are not valid sources or targets.

Output

{ "rows_written": 1200, "new_column": "hash_id_email", "output": "/data/customers.parquet" }

See also