Filtering & Sorting

Filter by exact values, ranges, or patterns. Sort ascending or descending on any column. Combine multiple conditions in a single request — all with plain query parameters.

What it does

Every csv-api dataset supports rich filtering and sorting through standard URL parameters. You can filter rows by exact match, comparison (greater than, less than), or pattern matching, and you can sort the results by any column in either direction. Filters and sorts compose freely, so you can ask for 'all customers in Portland over 30, sorted by signup date descending' in a single HTTP call.

How it works

1. 1

   Filter by column

   Pass any column name as a query parameter. ?city=Portland returns only rows where city equals Portland. Combine multiple filters by chaining them: ?city=Portland&status=active.

2. 2

   Use comparison operators

   Append _gt, _gte, _lt, _lte, or _ne to filter on numeric or date columns. ?age_gt=30&age_lte=50 returns rows where age is between 31 and 50.

3. 3

   Pattern matching for strings

   Use _like for substring search on text columns. ?name_like=ali matches Alice, Khali, and Mali.

4. 4

   Sort with one parameter

   ?sort=age sorts ascending. Prefix with - for descending: ?sort=-age. The same column you can filter, you can sort.

See it in action

# All Portland customers over 30, newest first, page 1
curl -H "Authorization: Bearer YOUR_API_KEY" \
  "https://csv-api.com/api/v1/datasets/d_a8f3bc91/records?\
city=Portland&age_gt=30&sort=-signup_date&per_page=20"

{
  "data": [
    { "name": "Alice", "age": 32, "city": "Portland", "signup_date": "2024-08-12" },
    { "name": "Carol", "age": 41, "city": "Portland", "signup_date": "2024-06-30" }
  ],
  "meta": { "total": 12, "page": 1, "per_page": 20, "total_pages": 1 }
}

Why it matters

• No query language to learn

  If you can build a URL, you can query your data. Standard query parameter syntax that works in any HTTP client.

• Operators respect column types

  Type detection means _gt and _lt work correctly on numbers and dates, while _like is reserved for strings. No coercion bugs.

• Composable filters

  Chain as many filter conditions as you want in a single request. They're combined with AND semantics.

Common use cases

• Powering search and filter UIs without writing a search backend

• Building dashboards that slice and dice the same dataset many ways

• Letting non-technical teammates pull subsets of data via URL

• Replacing one-off SQL queries with reusable, shareable URLs