Impute Missing Values

Overview

You send a rectangular table, the columns to fill, and a strategy: CONSTANT, MEAN, MEDIAN, or MOST_FREQUENT. CONSTANT requires constant and is the only way to fill an all-null column. MEAN and MEDIAN require INTEGER or NUMBER and promote INTEGER to NUMBER. A most-frequent tie keeps the first-seen value. The input table is copied. This is not interpolation, KNN, or a formula.

Filling nulls in place, or picking an arbitrary mode on a tie, makes a later checksum undebuggable. The table is copied and the fill value is reported.

When to use it

Scope

Non-goals

• Does not interpolate or run KNN
• Does not group before filling
• Does not evaluate formulas

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.impute_missing_values","input":{"table":{"columns":[{"name":"n","type":"INTEGER"}],"rows":[[1], [null], [3]]},"columns":["n"],"strategy":"MEAN"}}'

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

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

const { output } = await sq.execute("data.impute_missing_values", {"table":{"rows":[[1],[null],[3]],"columns":[{"type":"INTEGER","name":"n"}]},"strategy":"MEAN","columns":["n"]});

Errors

Other limits

Limitations

• Does not interpolate or run KNN
• Does not group before filling
• Does not evaluate formulas
• Time-series interpolation
• Model-based imputation

How to call

POST /v1/execute (requires authentication)

{
  "skillKey": "data.impute_missing_values",
  "input": {
    "table": {
      "rows": [
        [
          1
        ],
        [
          null
        ],
        [
          3
        ]
      ],
      "columns": [
        {
          "type": "INTEGER",
          "name": "n"
        }
      ]
    },
    "strategy": "MEAN",
    "columns": [
      "n"
    ]
  }
}

Examples

{
  "table": {
    "rows": [
      [
        1
      ],
      [
        null
      ],
      [
        3
      ]
    ],
    "columns": [
      {
        "type": "INTEGER",
        "name": "n"
      }
    ]
  },
  "strategy": "MEAN",
  "columns": [
    "n"
  ]
}

{
  "rows": [
    [
      1
    ],
    [
      2
    ],
    [
      3
    ]
  ],
  "rowCount": 3,
  "strategy": "MEAN",
  "imputations": [
    {
      "imputedCount": 1,
      "fillValue": 2,
      "column": "n"
    }
  ],
  "columns": [
    {
      "type": "NUMBER",
      "name": "n"
    }
  ],
  "columnCount": 1
}

{
  "table": {
    "rows": [
      [
        "a"
      ],
      [
        "b"
      ],
      [
        "a"
      ],
      [
        "b"
      ],
      [
        null
      ]
    ],
    "columns": [
      {
        "type": "STRING",
        "name": "c"
      }
    ]
  },
  "strategy": "MOST_FREQUENT",
  "columns": [
    "c"
  ]
}

{
  "rows": [
    [
      "a"
    ],
    [
      "b"
    ],
    [
      "a"
    ],
    [
      "b"
    ],
    [
      "a"
    ]
  ],
  "rowCount": 5,
  "strategy": "MOST_FREQUENT",
  "imputations": [
    {
      "imputedCount": 1,
      "fillValue": "a",
      "column": "c"
    }
  ],
  "columns": [
    {
      "type": "STRING",
      "name": "c"
    }
  ],
  "columnCount": 1
}

Recommended use cases

• Fill a numeric amount with the column mean before a union
• Resolve a most-frequent tie by first appearance on a STRING column

Avoid for

• Time-series interpolation
• Model-based imputation

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