Filter Rows

Overview

When you send tabular rows and a filter expression, we return only the rows that match — using a safe, compiled language (comparisons, AND/OR/NOT, string checks, in/not-in, null tests). No scripting, no user-supplied regex, no reflection. Do not use this for full SQL semantics or arbitrary code execution.

Row filtering by an LLM is slow and inconsistent. You get the same matches every time for the same rows and expression.

When to use it

Scope

Non-goals

• Does not execute code, SQL, or JavaScript
• Does not sort or aggregate rows
• Does not fetch data from files or URLs
• Does not accept user-supplied regular expressions

Properties

Capabilities

Cost & billing

Flat rate per invocation

File handling

How we handle your data

Input schema

Output schema

API contract

curl -sS -X POST "$BASE/v1/execute" -H "Authorization: Bearer $KEY" -H 'Content-Type: application/json' -d '{"skillKey":"data.filter_rows","input":{"rows":[{"id":1,"country":"BE","amount":150},{"id":2,"country":"FR","amount":200}],"filter":"country = \"BE\" AND amount >= 100"}}'

import { Squillify } from "@squillify/sdk";

const sq = new Squillify({ apiKey: process.env.SQUILLIFY_API_KEY });

const { output } = await sq.execute("data.filter_rows", {"filter":"country = \"BE\" AND amount >= 100","rows":[{"amount":150,"id":1,"country":"BE"},{"amount":200,"id":2,"country":"FR"}]});

Errors

Other limits

Limitations

• Does not execute code, SQL, or JavaScript
• Does not sort or aggregate rows
• Does not fetch data from files or URLs
• Does not accept user-supplied regular expressions
• Full SQL join/group semantics
• Fuzzy or regex-from-input matching
• Filtering more than 100,000 rows in one call

How to call

POST /v1/execute (requires authentication)

{
  "skillKey": "data.filter_rows",
  "input": {
    "filter": "country = \"BE\" AND amount >= 100",
    "rows": [
      {
        "amount": 150,
        "id": 1,
        "country": "BE"
      },
      {
        "amount": 200,
        "id": 2,
        "country": "FR"
      }
    ]
  }
}

Examples

{
  "filter": "country = \"BE\" AND amount >= 100",
  "rows": [
    {
      "amount": 150,
      "id": 1,
      "country": "BE"
    },
    {
      "amount": 200,
      "id": 2,
      "country": "FR"
    }
  ]
}

{
  "ok": true,
  "matchedCount": 1,
  "scannedCount": 2,
  "truncated": false,
  "rows": [
    {
      "country": "BE",
      "id": 1,
      "amount": 150
    }
  ]
}

{
  "filter": "status in [\"open\", \"pending\"]",
  "rows": [
    {
      "status": "open",
      "id": 1
    },
    {
      "status": "closed",
      "id": 2
    }
  ]
}

{
  "ok": true,
  "matchedCount": 1,
  "scannedCount": 2,
  "truncated": false,
  "rows": [
    {
      "status": "open",
      "id": 1
    }
  ]
}

Recommended use cases

• Filter API or CSV rows before export
• Select rows before aggregation
• Apply a deterministic WHERE-style predicate in an agent pipeline

Avoid for

• Full SQL join/group semantics
• Fuzzy or regex-from-input matching
• Filtering more than 100,000 rows in one call

Related skills

Composable with

Safety & privacy

• Runs without calling any external service.
• Your payloads are never stored permanently.
• Telemetry captures metadata only — never your payload contents.
• We never train models on your payloads.
• No human reviews your payloads.

Tags