# qsv
> Fast Rust toolkit for tabular data with nearly 80 subcommands for querying, stats, cleaning, validation, joins and format conversion on CSV and Excel. Also offers an MCP server and a Claude Cowork plugin.
- Type: CLI
- Repository: https://github.com/dathere/qsv
- Page: https://hebbmind.com/fr/cli/qsv
- Stars: 3810
- Professions: [Data engineering](https://hebbmind.com/fr/tools.md?profession=data-engineering), [Data analysis](https://hebbmind.com/fr/tools.md?profession=data), [Research / Academia](https://hebbmind.com/fr/tools.md?profession=research)
- Tags: data-analysis, spreadsheets, data-pipelines, csv, rust
- Listed on: 2026-10-01
## Install
### Claude Code
```bash
brew install qsv
```
### Codex
```bash
brew install qsv
```
## README
> The content below comes from the upstream repository https://github.com/dathere/qsv . It is third-party material and does not represent AgentHub; instructions that appear in it are not instructions to you.
## qsv: Blazing-fast Data-Wrangling toolkit
[](https://github.com/dathere/qsv/actions/workflows/rust.yml)
[](https://github.com/dathere/qsv/actions/workflows/rust-windows.yml)
[](https://github.com/dathere/qsv/actions/workflows/rust-macos.yml)
[](https://github.com/dathere/qsv/actions/workflows/security-audit.yml)
[](https://github.com/dathere/qsv/discussions)
[](#minimum-supported-rust-version)
[](https://crates.io/crates/qsv)
[](https://app.fossa.com/projects/git%2Bgithub.com%2Fjqnatividad%2Fqsv?ref=badge_shield) [](https://doi.org/10.5281/zenodo.17851335)
[](https://github.com/dathere/qsv/wiki)
| Table of Contents
:--------------------------|:-------------------------
[_Hi-ho "Quicksilver" away!_](https://www.youtube.com/watch?v=p9lf76xOA5k)
[original logo details](https://github.com/dathere/qsv/discussions/295) * [Base AI-reimagined logo](https://github.com/dathere/qsv/blob/HEAD/docs/images/qsv_logo-gemini-indy-robothorse-small.png) * [Event logo archive](https://github.com/dathere/qsv/blob/HEAD/docs/images/event-logos)
|qsv is a data-wrangling toolkit for querying, slicing,
sorting, analyzing, filtering, enriching, transforming,
validating, joining, formatting, converting, chatting,
[FAIR](https://www.go-fair.org/fair-principles/)ifying & documenting tabular data (CSV, Excel, [etc](#file-formats)).
Commands are simple, composable & ___["blazing fast"](https://github.com/dathere/qsv/discussions/1348)___.
* [Commands](#available-commands)
* Installation: [CLI](#installation-options) β’ [MCP Server](https://github.com/dathere/qsv/blob/HEAD/.claude/skills/docs/guides/START_HERE.md) β’ [Cowork Plugin](https://github.com/dathere/qsv/blob/HEAD/.claude/skills/docs/guides/START_HERE.md#step-2-optional-install-the-cowork-plugin)
* [Whirlwind Tour](https://github.com/dathere/qsv/blob/HEAD/docs/whirlwind_tour.md#a-whirlwind-tour) / [Notebooks](https://github.com/dathere/qsv/blob/HEAD/contrib/notebooks) / [Lessons & Exercises](https://100.dathere.com)
* [FAQ](https://github.com/dathere/qsv/discussions/categories/faq)
* [Performance Tuning](https://github.com/dathere/qsv/blob/HEAD/docs/PERFORMANCE.md#performance-tuning)
* π [Benchmarks](https://qsv.dathere.com/benchmarks) β’ [Interactive Benchmarks](https://github.com/dathere/qsv/wiki/Benchmarks) π
* π [Visualizations Gallery](https://dathere.github.io/qsv/gallery.html) πΊοΈ
* ποΈ [Neuro](https://dathere.github.io/qsv/smart_allegheny_dogs.html)-[symbolic](https://dathere.github.io/qsv/smart_boston_311_2025.html), [Localizable](https://dathere.github.io/qsv/smart-colombia-calidad-aire.html), _**[Data](https://dathere.github.io/qsv/smart_nyc311.html) [Schematics](https://dathere.github.io/qsv/pitt311data.html)**_ π
* [Environment Variables](https://github.com/dathere/qsv/blob/HEAD/docs/ENVIRONMENT_VARIABLES.md)
* [Feature Flags](#feature-flags)
* [Goals/Non-goals](#goals--non-goals)
* [Testing](#testing)
* [NYC SOD 2022](https://docs.google.com/presentation/d/e/2PACX-1vQ12ndZL--gkz0HLQRaxqsNOwzddkv1iUKB3sq661yA77OPlAsmHJHpjaqt9s9QEf73VqMfb0cv4jHU/pub?start=false&loop=false&delayms=3000)/[csv,conf,v8](https://docs.google.com/presentation/d/10T_3MyIqS5UsKxJaOY7Ktrd-GfhJelQImlE_qYmtuis/edit#slide=id.g2e0f1e7aa0e_0_62)/[PyConUS 2025](https://docs.google.com/presentation/d/e/2PACX-1vRKFnU0Hm8oDrtCYbxcf96kHVsPcoLU05jPVNYaAs09D05gPMWDJ96q_4_zgUvadGro4deohisy-XtY/pub?start=false&loop=false&delayms=3000)/
[csv,conf,v9](https://docs.google.com/presentation/d/1j-S0q5gqR8agsqIPBVXabGEntMlc4FDTwb4r-v8-9tA/edit?usp=sharing)/[NYC SOD 2026](https://docs.google.com/presentation/d/e/2PACX-1vTobPFucA1QO6u8dF3CyHjOctoom1DBmgQF558I_gx5e8cWPr0HLvJISvoaZyCMwLZZdDHlhK2cil0o/pub?start=false&loop=false&delayms=5000)
* **_"Have we achieved ACI?"_** series - [1](https://dathere.com/2026/01/the-peoples-api-is-finally-here/) β’ [2](https://dathere.github.io/NYC-Snow-Analysis-2010-2026-Claude-Cowork/) β’ [3](https://dathere.github.io/peoples-api-demos/NYC-Housing-Policy-SOD2026/) β’ [4](https://dathere.com/2026/08/data-schematic-a-neuro-symbolic-visual-story-telling-data-explorer-dictionary/)
* [Sponsor](#sponsor)
## Try it out at [qsv.dathere.com](https://qsv.dathere.com)!
| Command | Description |
| --- | --- |
| [apply](https://github.com/dathere/qsv/blob/HEAD/docs/help/apply.md)β¨
ππ§ π€ππ£πβ©οΈ | Apply series of string, date, math & currency transformations to given CSV column/s. It also has some basic [NLP](https://en.wikipedia.org/wiki/Natural_language_processing) functions ([similarity](https://crates.io/crates/strsim), [sentiment analysis](https://crates.io/crates/vader_sentiment), [profanity](https://docs.rs/censor/latest/censor/), [eudex](https://github.com/ticki/eudex#eudex-a-blazingly-fast-phonetic-reductionhashing-algorithm), [language](https://crates.io/crates/whatlang) & [name gender](https://github.com/Raduc4/gender_guesser?tab=readme-ov-file#gender-guesser)) detection. Its `summarize` subcommand condenses a column or group of columns using an OpenAI API-compatible LLM (local or commercial) with customizable, [MiniJinja](https://docs.rs/minijinja/latest/minijinja/)-templated per-record prompts. |
| [applydp](https://github.com/dathere/qsv/blob/HEAD/docs/help/applydp.md)β¨
πππ£π  | applydp is a slimmed-down version of `apply` with only [Datapusher+](https://github.com/dathere/datapusher-plus) relevant subcommands/operations (`qsvdp` binary variant only). |
| [behead](https://github.com/dathere/qsv/blob/HEAD/docs/help/behead.md) | Drop headers from a CSV. |
| [blake3](https://github.com/dathere/qsv/blob/HEAD/docs/help/blake3.md)
π | Compute or check [BLAKE3](https://github.com/BLAKE3-team/BLAKE3/?tab=readme-ov-file#blake3) hashes of files. |
| [cat](https://github.com/dathere/qsv/blob/HEAD/docs/help/cat.md)
ποΈ | Concatenate CSV files by row or by column. |
| [clean](https://github.com/dathere/qsv/blob/HEAD/docs/help/clean.md) | Remove qsv-generated cache files (`.idx` index, `stats` & `frequency` caches) to reduce clutter & simplify data packaging. With `--stale`, only removes caches whose source changed or is gone. Opt-in flags also clean `schema`, `validate` & `moarstats` outputs. |
| [clipboard](https://github.com/dathere/qsv/blob/HEAD/docs/help/clipboard.md)β¨
π₯οΈ | Provide input from the clipboard or save output to the clipboard. |
| [color](https://github.com/dathere/qsv/blob/HEAD/docs/help/color.md)β¨
π€―π»ββοΈπ₯οΈ | Outputs tabular data as a pretty, colorized table that always fits into the terminal. Apart from CSV and its dialects, Arrow, Avro/IPC, Parquet, JSON array & JSONL formats are supported with the "polars" feature. |
| [count](https://github.com/dathere/qsv/blob/HEAD/docs/help/count.md)
ππ»ββοΈποΈ | Count the rows and optionally compile record width statistics of a CSV file. (11.87 seconds for a 15gb, 28m row NYC 311 dataset without an index. Instantaneous with an index.) If the `polars` feature is enabled, uses Polars' multithreaded, mem-mapped CSV reader for fast counts even without an index |
| [datefmt](https://github.com/dathere/qsv/blob/HEAD/docs/help/datefmt.md)
πππ | Formats recognized date fields ([19 formats recognized](https://docs.rs/qsv-dateparser/latest/qsv_dateparser/#accepted-date-formats)) to a specified date format using [strftime date format specifiers](https://docs.rs/chrono/latest/chrono/format/strftime/). |
| [dedup](https://github.com/dathere/qsv/blob/HEAD/docs/help/dedup.md)
π€―ππ | Remove duplicate rows (See also `extdedup`, `extsort`, `sort` & `sortcheck` commands). |
| [denull](https://github.com/dathere/qsv/blob/HEAD/docs/help/denull.md) | Detect null sentinels β literal text like `NULL` or `N/A` standing in for a missing value, which makes `stats` type a numeric column as String (its `nullcount` stays 0, no quartiles are computed) and silently degrades `viz`, `schema` & `describegpt` downstream. Reports by default; `--apply` blanks the sentinels in the columns it confirmed, per column. Scans once with bounded memory. Numeric sentinels (`-999`) are deliberately NOT detected β they parse as valid numbers and no scan can tell them from real data. |
| [describegpt](https://github.com/dathere/qsv/blob/HEAD/docs/help/describegpt.md)
πποΈπ€ππͺπβ©οΈ  | Infer a ["neuro-symbolic"](https://en.wikipedia.org/wiki/Neuro-symbolic_AI) Data Dictionary, Description & Tags or ask questions about a CSV with a [configurable, MiniJinja prompt file](https://github.com/dathere/qsv/blob/HEAD/resources/describegpt_defaults.toml), using any [OpenAI API](https://platform.openai.com/docs/introduction)-compatible cloud/local LLM. (e.g. [Markdown](https://github.com/dathere/qsv/blob/HEAD/docs/describegpt/nyc311-describegpt.md), [JSON](https://github.com/dathere/qsv/blob/HEAD/docs/describegpt/nyc311-describegpt.json), [TOON](https://github.com/dathere/qsv/blob/HEAD/docs/describegpt/nyc311-describegpt.toon), [JSON Schema](https://github.com/dathere/qsv/blob/HEAD/docs/describegpt/nyc311-describegpt-schema.json), [Semantic Markdown](https://github.com/dathere/qsv/blob/HEAD/docs/describegpt/nyc311-describegpt-semanticmd.md), [OKF](https://github.com/dathere/qsv/blob/HEAD/docs/describegpt/nyc311-describegpt-okf.md), [Everything](https://github.com/dathere/qsv/blob/HEAD/docs/describegpt/nyc311-describegpt-everything.md), [Content Type inferencing](https://github.com/dathere/qsv/blob/HEAD/docs/describegpt/property-alleghenney.md), [Spanish](https://github.com/dathere/qsv/blob/HEAD/docs/describegpt/nyc311-describegpt-spanish.md), [Mandarin](https://github.com/dathere/qsv/blob/HEAD/docs/describegpt/nyc311-describegpt-mandarin.md), [Controlled Tags](https://github.com/dathere/qsv/blob/HEAD/docs/describegpt/nyc311-describegpt-tagvocab.md);
[--prompt "What are the top 10 complaint types by community board & borough by year?"](https://github.com/dathere/qsv/blob/HEAD/docs/describegpt/nyc311-describegpt-prompt.md) - [deterministic, hallucination-free SQL RAG result](https://github.com/dathere/qsv/blob/HEAD/docs/describegpt/nyc311-describegpt-prompt.csv); [iterative, session-based SQL RAG refinement](https://github.com/dathere/qsv/blob/HEAD/docs/describegpt/allegheny_discussion3.md) - [refined SQL RAG result](https://github.com/dathere/qsv/blob/HEAD/docs/describegpt/mostexpensive6.csv)) |
| [diff](https://github.com/dathere/qsv/blob/HEAD/docs/help/diff.md)
π | Find the difference between two CSVs with ludicrous speed!
e.g. _compare two CSVs with 1M rows x 9 columns in under 600ms!_ |
| [edit](https://github.com/dathere/qsv/blob/HEAD/docs/help/edit.md) | Replace the value of a cell specified by its row and column. |
| [enum](https://github.com/dathere/qsv/blob/HEAD/docs/help/enum.md)
π | Add a new column enumerating rows by adding a column of incremental or uuid identifiers. Can also be used to copy a column or fill a new column with a constant value. |
| [excel](https://github.com/dathere/qsv/blob/HEAD/docs/help/excel.md)
π | Exports a specified Excel/ODS sheet to a CSV file. |
| [exclude](https://github.com/dathere/qsv/blob/HEAD/docs/help/exclude.md)
π | Removes a set of CSV data from another set based on the specified columns. |
| [explode](https://github.com/dathere/qsv/blob/HEAD/docs/help/explode.md)
π£π | Explode rows into multiple ones by splitting a column value based on the given separator. The inverse of `implode`. |
| [extdedup](https://github.com/dathere/qsv/blob/HEAD/docs/help/extdedup.md)
π | Remove duplicate rows from an arbitrarily large CSV/text file using a memory-mapped, [on-disk hash table](https://crates.io/crates/odht). Unlike the `dedup` command, this command does not load the entire file into memory nor does it sort the deduped file. |
| [extsort](https://github.com/dathere/qsv/blob/HEAD/docs/help/extsort.md)
πππ | Sort an arbitrarily large CSV/text file using a multithreaded [external merge sort](https://en.wikipedia.org/wiki/External_sorting) algorithm. |
| [fetch](https://github.com/dathere/qsv/blob/HEAD/docs/help/fetch.md)β¨
ππ§ π | Send/Fetch data to/from web services for every row using **HTTP Get**. Comes with [HTTP/2](https://http2-explained.haxx.se/en/part1) [adaptive flow control](https://medium.com/coderscorner/http-2-flow-control-77e54f7fd518), [jaq](https://github.com/01mf02/jaq?tab=readme-ov-file#jaq) JSON query language support, dynamic throttling ([RateLimit](https://www.ietf.org/archive/id/draft-ietf-httpapi-ratelimit-headers-06.html)) & caching with available persistent caching using [Redis](https://redis.io/) or a disk-cache. |
| [fetchpost](https://github.com/dathere/qsv/blob/HEAD/docs/help/fetchpost.md)β¨
ππ§ πβ©οΈ | Similar to `fetch`, but uses **HTTP Post** ([HTTP GET vs POST methods](https://www.geeksforgeeks.org/difference-between-http-get-and-post-methods/)). Supports HTML form (application/x-www-form-urlencoded), JSON (application/json) and custom content types - with the ability to render payloads using CSV data using the [MiniJinja](https://docs.rs/minijinja/latest/minijinja/) template engine. |
| [fill](https://github.com/dathere/qsv/blob/HEAD/docs/help/fill.md)
π | Fill empty values. |
| [fixedwidth](https://github.com/dathere/qsv/blob/HEAD/docs/help/fixedwidth.md) | Convert fixed-width text (fields at fixed byte-column positions, no delimiters) to CSV. |
| [fixlengths](https://github.com/dathere/qsv/blob/HEAD/docs/help/fixlengths.md) | Force a CSV to have same-length records by either padding or truncating them. |
| [flatten](https://github.com/dathere/qsv/blob/HEAD/docs/help/flatten.md) | A flattened view of CSV records. Useful for viewing one record at a time.
e.g. `qsv slice -i 5 data.csv \| qsv flatten`. |
| [fmt](https://github.com/dathere/qsv/blob/HEAD/docs/help/fmt.md) | Reformat a CSV with different delimiters, record terminators or quoting rules. (Supports ASCII delimited data.) |
| [foreach](https://github.com/dathere/qsv/blob/HEAD/docs/help/foreach.md)β¨ | Execute a shell command once per record in a given CSV file. |
| [frequency](https://github.com/dathere/qsv/blob/HEAD/docs/help/frequency.md)
ππ£ποΈππͺ | Build [frequency distribution tables](https://en.wikipedia.org/wiki/Frequency_(statistics)) of each column. Uses multithreading to go faster if an index is present (Examples: [CSV](https://github.com/dathere/qsv/blob/HEAD/scripts/nyc311-1m.freqs.csv) [JSON](https://github.com/dathere/qsv/blob/HEAD/scripts/nyc311-1m.freqs.json) [TOON](https://github.com/dathere/qsv/blob/HEAD/scripts/nyc311-1m.freqs.toon)). |
| [get](https://github.com/dathere/qsv/blob/HEAD/docs/help/get.md)β¨
ππ§ π  | Get tabular data from local files, URLs (http/https & `dathere://`) & [CKAN](https://ckan.org) (`ckan://`) into a managed, queryable disk cache - with conditional revalidation (ETag/Last-Modified), transparent [zstd](https://github.com/facebook/zstd) compression, [BLAKE3](https://github.com/BLAKE3-team/BLAKE3) hashing & automatic indexing. Cached resources are reusable by ANY qsv command via the `dc:` prefix (e.g. `qsv stats dc:data.csv`), with stale entries auto-refreshed. Efficiently seeds `luau` lookup tables, `validate` dynamicEnum reference data & speeds up Datapusher+ harvesting. |
| [geocode](https://github.com/dathere/qsv/blob/HEAD/docs/help/geocode.md)β¨
ππ§ πππ£ππ | Geocodes a location against an updatable local copy of the [Geonames](https://www.geonames.org/) cities & the [Maxmind GeoLite2](https://www.maxmind.com/en/geolite-free-ip-geolocation-data) databases β with caching and multi-threading, this offline path geocodes up to 360,000 records/sec! Can also geocode online (forward & reverse) via the [OpenCage](https://opencagedata.com) geocoder, geocode New York City addresses, BBLs and BINs via NYC's [Geoclient](https://github.com/mlipper/geoclient) API, which also returns NYC attributes such as council districts and census tracts, and geocode US addresses & coordinates via the free [US Census Geocoder](https://geocoding.geo.census.gov/geocoder/) - single or batch - returning Census state, county, tract, block group & block GEOIDs for Census data enrichment. |
| [geoconvert](https://github.com/dathere/qsv/blob/HEAD/docs/help/geoconvert.md)β¨
π | Convert between various spatial formats and CSV/SVG including GeoJSON, SHP, and more. |
| [headers](https://github.com/dathere/qsv/blob/HEAD/docs/help/headers.md)
ποΈ | Show the headers of a CSV. Or show the intersection of all headers between many CSV files. |
| [implode](https://github.com/dathere/qsv/blob/HEAD/docs/help/implode.md)
π£π | Implode rows by grouping on key column(s) and joining a value column with a given separator. The inverse of `explode`. |
| [index](https://github.com/dathere/qsv/blob/HEAD/docs/help/index.md) | Create an index (π) for a CSV. This is very quick (even the 15gb, 28m row NYC 311 dataset takes all of 14 seconds to index) & provides constant time indexing/random access into the CSV. With an index, `count`, `sample` & `slice` work instantaneously; random access mode is enabled in `luau`; and multithreading (ποΈ) is enabled for the `frequency`, `split`, `stats` & `schema` commands. |
| [input](https://github.com/dathere/qsv/blob/HEAD/docs/help/input.md) | Read CSV data with special commenting, quoting, trimming, line-skipping & non-UTF8 encoding handling rules. Typically used to "normalize" a CSV for further processing with other qsv commands. |
| [join](https://github.com/dathere/qsv/blob/HEAD/docs/help/join.md)
ππ£π | Inner, outer, right, cross, anti & semi joins. Automatically creates a simple, in-memory hash index to make it fast. |
| [joinp](https://github.com/dathere/qsv/blob/HEAD/docs/help/joinp.md)β¨
π»ββοΈππͺ | Inner, outer, right, cross, anti, semi, non-equi & asof joins using the [Pola.rs](https://www.pola.rs) engine. Unlike the `join` command, `joinp` can process files larger than RAM, is multithreaded, has join key validation, a maintain row order option, pre and post-join filtering, join keys unicode normalization, supports "special" [non-equi joins](https://docs.pola.rs/user-guide/transformations/joins/#non-equi-joins) and [asof joins](https://docs.pola.rs/user-guide/transformations/joins/#asof-join) (which is [particularly useful for time series data](https://github.com/dathere/qsv/blob/30cc920d0812a854fcbfedc5db81788a0600c92b/tests/test_joinp.rs#L509-L983)) & its output columns can be coalesced. |
| [json](https://github.com/dathere/qsv/blob/HEAD/docs/help/json.md)
π | Convert JSON array to CSV.
| [jsonl](https://github.com/dathere/qsv/blob/HEAD/docs/help/jsonl.md)
ππ£ | Convert newline-delimited JSON ([JSONL](https://jsonlines.org/)/[NDJSON](http://ndjson.org/)) to CSV. See `tojsonl` command to convert CSV to JSONL.
| [lens](https://github.com/dathere/qsv/blob/HEAD/docs/help/lens.md)β¨
ποΈπ»ββοΈπ₯οΈ | Interactively view, search & filter tabular data files using the [csvlens](https://github.com/YS-L/csvlens#csvlens) engine. Apart from CSV and its dialects, Arrow, Avro/IPC, Parquet, JSON array & JSONL formats are supported with the "polars" feature. |
| [luau](https://github.com/dathere/qsv/blob/HEAD/docs/help/luau.md)β¨
πππ£π   | Create multiple new computed columns, filter rows, compute aggregations and build complex data pipelines by executing a [Luau](https://luau-lang.org) [0.740](https://github.com/Roblox/luau/releases/tag/0.740) expression/script for every row of a CSV file ([sequential mode](https://github.com/dathere/qsv/blob/bb72c4ef369d192d85d8b7cc6e972c1b7df77635/tests/test_luau.rs#L254-L298)), or using [random access](https://www.webopedia.com/definitions/random-access/) with an index ([random access mode](https://github.com/dathere/qsv/blob/bb72c4ef369d192d85d8b7cc6e972c1b7df77635/tests/test_luau.rs#L367-L415)).
Can process a single Luau expression or [full-fledged data-wrangling scripts using lookup tables](https://github.com/dathere/qsv-lookup-tables#example) with discrete BEGIN, MAIN and END sections.
It is not just another qsv command, it is qsv's [Domain-specific Language](https://en.wikipedia.org/wiki/Domain-specific_language) (DSL) with [numerous qsv-specific helper functions](https://github.com/dathere/qsv/blob/master/src/cmd/luau.rs#L1473-L2755) to build production data pipelines. |
| [moarstats](https://github.com/dathere/qsv/blob/HEAD/docs/help/moarstats.md)
πποΈ | Add up to an additional 73 statistical measures, including extended outlier, robust & bivariate statistics to an existing stats CSV file. ([example](https://github.com/dathere/qsv/blob/HEAD/docs/moarstats/NYC_311_SR_2010-2020-sample-1M.stats.csv)).|
| [partition](https://github.com/dathere/qsv/blob/HEAD/docs/help/partition.md)
π | Partition a CSV based on a column value. |
| [pivotp](https://github.com/dathere/qsv/blob/HEAD/docs/help/pivotp.md)β¨
π»ββοΈππͺ | Pivot CSV data. Features "smart" aggregation auto-selection based on data type & stats. |
| [pragmastat](https://github.com/dathere/qsv/blob/HEAD/docs/help/pragmastat.md)
ππ€―π²πͺ | Compute pragmatic statistics using the [Pragmastat](https://pragmastat.dev/) library. Uses the stats cache to auto-filter non-numeric columns and support Date/DateTime columns. |
| [pro](https://github.com/dathere/qsv/blob/HEAD/docs/help/pro.md) | Interact with the [qsv pro](https://qsvpro.dathere.com) API. |
| [profile](https://github.com/dathere/qsv/blob/HEAD/docs/help/profile.md)β¨
ππ§ π€πβ©οΈ  | Extract, derive & infer metadata from a CSV (local path or URL) - using the statistical profile of a dataset, mapped and driven by a configurable metadata [scheming](https://github.com/ckan/ckanext-scheming) YAML spec ([DCAT-US v3](https://resources.data.gov/resources/dcat-us3/), [DCAT-AP v3](https://semiceu.github.io/DCAT-AP/releases/3.0.0/) and [Croissant 1.1](https://docs.mlcommons.org/croissant/docs/croissant-spec-1.1.html) bundled; [Geoconnex](https://docs.geoconnex.us/reference/overview) when built with the `geoconnex` feature), with optional CKAN/DCAT metadata discovery for URL inputs. This enables [FAIRification](https://www.go-fair.org/fair-principles/fairification-process/) at scale. |
| [prompt](https://github.com/dathere/qsv/blob/HEAD/docs/help/prompt.md)β¨
π»ββοΈπ₯οΈ | Open a file dialog to either pick a file as input or save output to a file. |
| [pseudo](https://github.com/dathere/qsv/blob/HEAD/docs/help/pseudo.md)
π£π | [Pseudonymise](https://en.wikipedia.org/wiki/Pseudonymization) the value of the given column by replacing them with an incremental identifier. |
| [py](https://github.com/dathere/qsv/blob/HEAD/docs/help/py.md)β¨
ππ£ | Create a new computed column or filter rows by evaluating a Python expression on every row of a CSV file. Python's [f-strings](https://www.freecodecamp.org/news/python-f-strings-tutorial-how-to-use-f-strings-for-string-formatting/) is particularly useful for extended formatting, [with the ability to evaluate Python expressions as well](https://github.com/dathere/qsv/blob/4cd00dca88addf0d287247fa27d40563b6d46985/src/cmd/python.rs#L23-L31). [Requires Python 3.11 or greater](https://github.com/dathere/qsv/blob/master/docs/INTERPRETERS.md#building-qsv-with-python-feature). |
| [readstat](https://github.com/dathere/qsv/blob/HEAD/docs/help/readstat.md)β¨
π€―π»ββοΈπ | Convert SAS (`.sas7bdat`, `.xpt`), Stata (`.dta`) & SPSS (`.sav`, `.zsav`, `.por`) files to CSV, preserving the underlying codes of labelled values by default. Also dumps the rich variable metadata these formats carry - variable labels, value labels, missing-value codes, measure & display settings - with `--metadata`, or writes them alongside the data as a JSON Schema data dictionary (`--dictionary`) that `validate` and `viz smart --dictionary` read, no LLM needed. Can read just the variables you need (`--select`) and a window or reproducible random sample of rows (`--offset`, `--limit`, `--sample`).
Only SPSS portable (`.por`) files are read whole - every other format streams with constant memory, except SPSS `.sav`/`.zsav` files whose user-defined missing values are tracked (`--sentinels-as`, or the default check that warns when they are dropped). |
| [rename](https://github.com/dathere/qsv/blob/HEAD/docs/help/rename.md) | Rename the columns of a CSV efficiently. |
| [replace](https://github.com/dathere/qsv/blob/HEAD/docs/help/replace.md)
πποΈπ | Replace CSV data using a regex. Applies the regex to each field individually. |
| [reverse](https://github.com/dathere/qsv/blob/HEAD/docs/help/reverse.md)
ππ€― | Reverse order of rows in a CSV. Unlike the `sort --reverse` command, it preserves the order of rows with the same key. If an index is present, it works with constant memory. Otherwise, it will load all the data into memory. |
| [safenames](https://github.com/dathere/qsv/blob/HEAD/docs/help/safenames.md)
 | Modify headers of a CSV to only have ["safe" names](https://github.com/dathere/qsv/blob/HEAD/src/cmd/safenames.rs#L5-L14) - guaranteed "database-ready"/"CKAN-ready" names. |
| [sample](https://github.com/dathere/qsv/blob/HEAD/docs/help/sample.md)
πποΈππ²πͺ | Randomly draw rows (with optional seed) from a CSV using ten different sampling methods - [reservoir](https://en.wikipedia.org/wiki/Reservoir_sampling) (default), [indexed](https://en.wikipedia.org/wiki/Random_access), [bernoulli](https://en.wikipedia.org/wiki/Bernoulli_sampling), [systematic](https://en.wikipedia.org/wiki/Systematic_sampling), [stratified](https://en.wikipedia.org/wiki/Stratified_sampling), [weighted](https://doi.org/10.1016/j.ipl.2005.11.003), [varopt](https://doi.org/10.1016/j.ipl.2005.11.003), [mergeable-reservoir](https://dl.acm.org/doi/10.1145/3147.3165), [cluster](https://en.wikipedia.org/wiki/Cluster_sampling) & timeseries sampling. The `--varopt` & `--mergeable-reservoir` modes support mergeable sketch I/O (`--sketch-out`/`--sketch-in`) so sharded inputs can be sampled and combined without re-reading the corpus. Supports sampling from CSVs on remote URLs. Uses the stats cache to skip unnecessary scanning and inform its sampling strategies. |
| [schema](https://github.com/dathere/qsv/blob/HEAD/docs/help/schema.md)
ππ£π»ββοΈποΈππͺ | Infer either a [JSON Schema Validation Draft 2020-12](https://json-schema.org/draft/2020-12/json-schema-validation) ([Example](https://github.com/dathere/qsv/blob/master/resources/test/311_Service_Requests_from_2010_to_Present-2022-03-04.csv.schema.json)) or [Polars Schema](https://docs.pola.rs/user-guide/lazy/schemas/) ([Example](https://github.com/dathere/qsv/blob/master/resources/test/NYC_311_SR_2010-2020-sample-1M.pschema.json)) from CSV data.
In JSON Schema Validation mode, it produces a `.schema.json` file replete with inferred data type & domain/range validation rules derived from [`stats`](#stats_deeplink). Uses multithreading to go faster if an index is present. See [`validate`](#validate_deeplink) command to use the generated JSON Schema to validate if similar CSVs comply with the schema.
With the `--polars` option, it produces a `.pschema.json` file that all polars commands (`sqlp`, `joinp` & `pivotp`) use to determine the data type of each column & to optimize performance.
Both schemas are editable and can be fine-tuned. For JSON Schema, to refine the inferred validation rules. For Polars Schema, to change the inferred Polars data types. |
| [scoresql](https://github.com/dathere/qsv/blob/HEAD/docs/help/scoresql.md)β¨
π»ββοΈπͺ | Analyze a SQL query against CSV file caches (stats, moarstats, frequency) to produce a performance score with actionable optimization suggestions BEFORE running the query. Supports Polars (default) and DuckDB modes. |
| [search](https://github.com/dathere/qsv/blob/HEAD/docs/help/search.md)
πποΈπ | Run a regex over a CSV. Applies the regex to selected fields & shows only matching rows. |
| [searchset](https://github.com/dathere/qsv/blob/HEAD/docs/help/searchset.md)
πποΈπ | _Run multiple regexes over a CSV in a single pass._ Applies the regexes to each field individually & shows only matching rows. |
| [select](https://github.com/dathere/qsv/blob/HEAD/docs/help/select.md)
π | Select, re-order, reverse, duplicate or drop columns. |
| [slice](https://github.com/dathere/qsv/blob/HEAD/docs/help/slice.md)
πποΈποΈ | Slice rows from any part of a CSV. When an index is present, this only has to parse the rows in the slice (instead of all rows leading up to the start of the slice). |
| [snappy](https://github.com/dathere/qsv/blob/HEAD/docs/help/snappy.md)
ππ | Does streaming compression/decompression of the input using Google's [Snappy](https://github.com/google/snappy/blob/main/docs/README.md) framing format ([more info](#automatic-compressiondecompression)). |
| [sniff](https://github.com/dathere/qsv/blob/HEAD/docs/help/sniff.md)
ππ€π  | Quickly sniff & infer CSV metadata (delimiter, header row, preamble rows, quote character, flexible, is_utf8, average record length, number of records, content length & estimated number of records if sniffing a CSV on a URL, number of fields, field names & data types). It is also a general mime type detector. |
| [sort](https://github.com/dathere/qsv/blob/HEAD/docs/help/sort.md)
π€―πππ² | Sorts CSV data in [lexicographical](https://en.wikipedia.org/wiki/Lexicographic_order), [natural](https://en.wikipedia.org/wiki/Natural_sort_order), numerical, reverse, unique or random (with optional seed) order (Also see `extsort` & `sortcheck` commands). |
| [sortcheck](https://github.com/dathere/qsv/blob/HEAD/docs/help/sortcheck.md)
π | Check if a CSV is sorted. With the --json options, also retrieve record count, sort breaks & duplicate count. |
| [split](https://github.com/dathere/qsv/blob/HEAD/docs/help/split.md)
πποΈ | Split one CSV file into many CSV files. It can split by number of rows, number of chunks or file size. Uses multithreading to go faster if an index is present when splitting by rows or chunks. |
| [sqlp](https://github.com/dathere/qsv/blob/HEAD/docs/help/sqlp.md)β¨
πποΈπ»ββοΈππͺ | Run [Polars](https://pola.rs) SQL (a PostgreSQL dialect) queries against several CSVs, Parquet, JSONL and Arrow files - converting queries to blazing-fast Polars [LazyFrame](https://docs.pola.rs/user-guide/lazy/) expressions, processing larger than memory CSV files. Query results can be saved in CSV, JSON, JSONL, Parquet, Apache Arrow IPC and Apache Avro formats. |
| [stats](https://github.com/dathere/qsv/blob/HEAD/docs/help/stats.md)
ππ€―ποΈππͺ | Compute up to 48 [summary statistics](https://github.com/dathere/qsv/blob/HEAD/docs/STATS_DEFINITIONS.md) & make GUARANTEED data type inferences (Null, String, Float, Integer, Date, DateTime, Boolean) for each column in a CSV ([Example](https://github.com/dathere/qsv/blob/master/docs/stats/NYC_311_SR_2010-2020-sample-1M.stats.csv)).
Uses multithreading to go faster if an index is present. With an index, can compile "streaming" stats on a 1M row sample of NYC's 311 data in [less than 0.25 seconds vs 2.24 seconds without one](https://qsv.dathere.com/benchmarks). |
| [synthesize](https://github.com/dathere/qsv/blob/HEAD/docs/help/synthesize.md)β¨
ππ²π€ | Generate a synthetic CSV that is statistically faithful to a source CSV. Runs `stats` + `frequency` on the source so synthesized columns reproduce its per-column attributes β frequency-weighted sampling for categorical columns, quartile-bucketed numeric/date generation, null-ratio preservation. With a Data Dictionary from `describegpt --dictionary --infer-content-type`, semantic Content Types pick realistic [fake-rs](https://github.com/cksac/fake-rs) fakers (names, emails, addresses, UUIDs, etc.) for non-enumerable columns. A dictionary `relationships` array preserves inter-column structure within each row β `joint` (functional dependencies like city/state/zip), `ordered` (monotonic chains like created_date β€ closed_date) and `correlated` (numeric correlation via a Gaussian copula). Fully reproducible with `--seed`. |
| [table](https://github.com/dathere/qsv/blob/HEAD/docs/help/table.md)
π€― | Align output of a CSV using [elastic tabstops](https://github.com/BurntSushi/tabwriter) for viewing; or to create an "aligned TSV" file or Fixed Width Format file. To interactively view a CSV, use the `lens` command. |
| [template](https://github.com/dathere/qsv/blob/HEAD/docs/help/template.md)
πππ£πβ©οΈ  | Renders a template using CSV data with the [MiniJinja](https://docs.rs/minijinja/latest/minijinja/) template engine ([Example](https://github.com/dathere/qsv/blob/4645ec07b5befe3b0c0e49bf0f547315d0d7514b/src/cmd/template.rs#L18-L44)). |
| [to](https://github.com/dathere/qsv/blob/HEAD/docs/help/to.md)β¨
ποΈπ»ββοΈπ | Convert CSV files to [Parquet](https://parquet.apache.org), [PostgreSQL](https://www.postgresql.org), [SQLite](https://www.sqlite.org/index.html), Excel (XLSX), [LibreOffice Calc](https://www.libreoffice.org/discover/calc/) (ODS) and [Data Package](https://datahub.io/docs/data-packages/tabular). |
| [tojsonl](https://github.com/dathere/qsv/blob/HEAD/docs/help/tojsonl.md)
ππ£ποΈππ£πͺ | Smartly converts CSV to a newline-delimited JSON ([JSONL](https://jsonlines.org/)/[NDJSON](http://ndjson.org/)). By scanning the CSV first, it "smartly" infers the appropriate JSON data type for each column. See `jsonl` command to convert JSONL to CSV. |
| [transpose](https://github.com/dathere/qsv/blob/HEAD/docs/help/transpose.md)
π€―π | Transpose rows/columns of a CSV. |
| [validate](https://github.com/dathere/qsv/blob/HEAD/docs/help/validate.md)
πποΈπππ  | Validate CSV data [_blazingly-fast_](https://github.com/Stranger6667/jsonschema-rs?tab=readme-ov-file#performance "using jsonschema-rs - the fastest JSON Schema validator for Rust") using [JSON Schema Validation (Draft 2020-12)](https://json-schema.org/draft/2020-12/json-schema-validation.html) (e.g. _up to 1,338,688 rows/sec_[^1] using [NYC's 311 schema](https://github.com/dathere/qsv/blob/master/resources/test/311_Service_Requests_from_2010_to_Present-2022-03-04.csv.schema.json) generated by the [`schema`](#schema_deeplink) command) & put invalid records into a separate file along with a detailed validation error report.
Supports several custom JSON Schema formats & keywords:
* `currency` custom format with [ISO-4217](https://en.wikipedia.org/wiki/ISO_4217) validation
* `dynamicEnum` custom keyword that supports enum validation against a CSV on the filesystem or a URL (http/https/ckan & dathere URL schemes supported)
* `uniqueCombinedWith` custom keyword to validate uniqueness across multiple columns for composite key validation.
If no JSON schema file is provided, validates if a CSV conforms to the [RFC 4180 standard](#rfc-4180-csv-standard) and is UTF-8 encoded. |
| [viz](https://github.com/dathere/qsv/blob/HEAD/docs/help/viz.md)β¨
πͺπποΈππ€ππ | Generate interactive charts & maps from CSV data using [plotly](https://plotly.com). `viz smart` creates a [Data Schematic](https://github.com/dathere/qsv/blob/HEAD/docs/DATA_SCHEMATIC.md) β a *"[neuro-symbolic](https://en.wikipedia.org/wiki/Neuro-symbolic_AI)"* interactive rendering of a dataset's schema & statistics β picking appropriate visualizations using the dataset's statistics, frequency distributions, data dictionary & optional LLM metadata inferencing/classification, with automatic geocoding enrichment. Outputs self-contained, interactive HTML or static PNG/SVG/PDF/JPEG/WebP with the `viz_static` feature. ([Gallery](https://dathere.github.io/qsv/gallery.html)) |
| [writestat](https://github.com/dathere/qsv/blob/HEAD/docs/help/writestat.md)β¨
π€―π»ββοΈ | The inverse of `readstat`: write a CSV as an SPSS (`.sav`, `.por`), Stata (`.dta`) or SAS transport (`.xpt`) file. With the JSON Schema data dictionary `readstat --dictionary` wrote, a converted file comes back with its variable labels, value labels, missing values & display settings - and decoded value labels & sentinel columns are turned back into codes & declared missing values. Metadata the output format can't hold is refused, not silently dropped, unless `--lossy` is given. |
Performance metrics compiled on an M2 Pro 12-core Mac Mini with 32gb RAM
β¨: enabled by a [feature flag](#feature-flags).
π: uses an index when available.
π€―: loads entire CSV into memory, though `dedup`, `stats` & `transpose` have "streaming" modes as well.
π£: uses additional memory proportional to the cardinality of the columns in the CSV.
π§ : expensive operations are memoized with available inter-session Redis/Disk caching for fetch commands.
ποΈ: [Extended input support](#extended-input-support).
ποΈ: [Limited Extended input support](#limited-extended-input-support).
π»ββοΈ: command powered/accelerated by [](https://github.com/pola-rs/polars/releases/tag/py-2.0.0) vectorized query engine.
π€: command uses Natural Language Processing or Generative AI.
ποΈ: multithreaded and/or faster when an index (π) is available.
π: multithreaded even without an index.
 : has [CKAN](https://ckan.org)-aware integration options.
π: has web-aware options.
π£: requires UTF-8 encoded input.
> The README is long; the above is its first 40000 characters. Full content: https://github.com/dathere/qsv