drop_column

Removes a column from a table, along with its values in every row.


Scope	`mcp:write`
Annotations	`destructiveHint: true`
Activity event	`column.dropped`

The column’s data can’t be recovered. Export the table first if you might need it. See Export to CSV.

Parameters

string

required

Project id from list_projects. You need write access.

string

required

Table name.

string

required

The column to drop. Can’t be id or created_at.

Result

The table’s updated definition. Same shape as the describe_table result.

Example

Arguments

{ "projectId": "01k7m3x9q2v8d4h6n1r5t0w2yz", "table": "books", "column": "details" }

Result

{
  "data": {
    "table": "books",
    "columns": [
      { "name": "id", "type": "uuid", "nullable": false, "default": "gen_random_uuid()" },
      { "name": "created_at", "type": "timestamptz", "nullable": false, "default": "now()" },
      { "name": "title", "type": "text", "nullable": false },
      { "name": "rating", "type": "integer" }
    ]
  }
}

Errors

Code	When
`ACCESS_DENIED`	You don’t have write access, or the connection is read-only
`VALIDATION`	Unknown table or column, a system column, or the table’s last remaining column of its own

Notes

If the column has a foreign key, the foreign key is removed with it.
A table must keep at least one column besides id and created_at. To remove the last one, use drop_table.

Overview

Read tools

Write tools

Destructive tools

Parameters

Result

Example

Errors

Notes

​Parameters

​Result

​Example

​Errors

​Notes

Parameters

Result

Example

Errors

Notes