Data Convert

Documents

data-convert

Data Convert is an Agent Skill for documents published in the dathere/qsv repository on GitHub: a SKILL.md instruction file an AI agent can load to do this task.

Convert between CSV, TSV, Excel, JSONL, Parquet, and other tabular formats

View on GitHub ↗

QUICK START

How to use this skill

Bring this guide into your coding agent with a prompt tailored to the tool you use.

Open your project in Codex.
Copy and paste the prompt below.
Review files and risks before approving any action.

Prompt to paste

Inspect the Agent Skill at https://github.com/dathere/qsv/blob/HEAD/.claude/skills/skills/data-convert/SKILL.md and all supporting files. Report provenance, permissions, scripts, dependencies, and security risks. Propose an exact installation into .agents/skills/data-convert, but wait for my approval before writing files or running scripts.

Copying this prompt does not install or run the skill. Review third-party files before use.

Data Convert

Convert tabular data files between formats.

Cowork note: If relative paths don't resolve, call mcp__qsv__qsv_get_working_dir and mcp__qsv__qsv_set_working_dir to sync the working directory.

Supported Conversions

Input Formats (auto-detected)

CSV (.csv), TSV (.tsv/.tab), SSV (.ssv)
Excel (.xlsx, .xls, .xlsm, .xlsb)
OpenDocument (.ods)
JSONL/NDJSON (.jsonl, .ndjson)
Snappy-compressed variants (.csv.sz, etc.)

Output Formats

Format	Command	Extension
CSV	`select` (identity) or `fmt`	`.csv`
TSV	`fmt --out-delimiter '\t'`	`.tsv`
JSONL	`tojsonl`	`.jsonl`
JSON	`slice --json`	`.json`
Parquet	`mcp__qsv__qsv_to_parquet` (core tool)	`.parquet`
XLSX	`to xlsx` (via `mcp__qsv__qsv_command`)	`.xlsx`
ODS	`to ods` (via `mcp__qsv__qsv_command`)	`.ods`
SQLite	`to sqlite` (via `mcp__qsv__qsv_command`)	`.db`
PostgreSQL	`to postgres` (via `mcp__qsv__qsv_command`)	N/A
Data Package	`to datapackage` (via `mcp__qsv__qsv_command`)	`.json`

Steps

Index: Run mcp__qsv__qsv_index on the file for fast random access in subsequent steps.
Detect source format: Run mcp__qsv__qsv_sniff to identify the input format, delimiter, and encoding.
Convert: Use the appropriate command based on the target format:
- To CSV (from Excel/JSONL): The MCP server handles this automatically when you pass non-CSV files to any qsv tool. Use mcp__qsv__qsv_command with excel for explicit control over sheet selection.
- To TSV: Use mcp__qsv__qsv_command with command: "fmt", options: {"out-delimiter": "\t"}.
- To JSONL: Use mcp__qsv__qsv_command with command: "tojsonl".
- To Parquet (single file): Use mcp__qsv__qsv_to_parquet (core tool) — auto-generates stats cache and Polars schema for optimal type inference.
- To Parquet (batch): Use mcp__qsv__qsv_command with command: "to", subcommand: "parquet", destination: "output_dir" for batch conversion with explicit compression control.
- To XLSX: Use mcp__qsv__qsv_command with command: "to", subcommand: "xlsx", destination: "output.xlsx".
- To ODS: Use mcp__qsv__qsv_command with command: "to", subcommand: "ods", destination: "output.ods".
- To SQLite: Use mcp__qsv__qsv_command with command: "to", subcommand: "sqlite", destination: "output.db".
- To PostgreSQL: Use mcp__qsv__qsv_command with command: "to", subcommand: "postgres", destination: "connection_string".
- To Data Package: Use mcp__qsv__qsv_command with command: "to", subcommand: "datapackage", destination: "output.json".
Verify output: Run mcp__qsv__qsv_count on the output (if CSV-based) to confirm row count matches input.

Notes

Excel conversion: Use --sheet to specify which sheet to convert (default: first sheet)
JSONL output respects data types from stats cache - run stats --stats-jsonl first for better type inference
Parquet conversion preserves data types efficiently and produces smaller files
For CSV -> CSV reformatting (change delimiter, quoting), use fmt command
Large Excel files may take longer to convert - the MCP server handles this transparently
When converting multiple sheets from Excel, run excel with --sheet for each sheet