csv-query

Query CSV data with structured filter, sort, select, and aggregate operations.

Why install it

CSV handling shows up constantly in ops and analyst workflows. This tool gives agents a safe, predictable query surface without embedding ad hoc dataframe code into each app.

Inputs

• path: path to a local CSV file
• csv_text: raw CSV content passed directly to the tool
• select: subset of columns to keep in the output
• filter: filter conditions such as eq, gt, or contains
• sort: sort operations with column and optional direction
• limit: maximum number of rows to return
• group_by: columns used to group rows before aggregation
• aggregations: aggregate calculations such as count, sum, avg, min, and max

Outputs

• columns: output column names
• rows: resulting rows after query operations
• row_count: number of rows returned
• summary: aggregate or grouping summary information

Local development

The source code for this tool can be found here

Test:

python -m unittest discover -s tests -p 'test_*.py'

Example invocation

python -u csv_query/__main__.py < input.json

With input.json containing:

 {
    "csv_text": "name,score\nA,2\nB,5",
    "filter": [
      {
        "column": "score",
        "op": "gt",
        "value": "2"
      }
    ]
  }