← All apps

#Datalake #GotSimplified

Data lake: queries, row changes, table structure and sharing (datalake_query, datalake_mutate, datalake_manage, datalake_share)

New to MCP? Learn how to connect this server in our Setup Guide → (Cursor, Claude, OpenCode, Windsurf, Goose).

What you can do

Create, move, change or drop data lake tables and roll them back to an earlier version.

Parameters for this action
  • namespacestringrequired
    The namespace (folder) of the table, e.g. 'hr'. Letters of any language, digits and _
  • table_namestringoptional
    The table to work on
  • columnsobjectoptional
    Create table: the columns and their types, e.g. id as a number and name as text
  • partition_byarray of stringoptional
    New tables: how to split the data for faster reads, e.g. by month or region (up to 3)
Example prompts
  • >Create a table for orders with id, customer and amount
  • >Create a sales table partitioned by month of order_date and region

Works with every action: wait (False (default): start the job and return a job id to check later. True: wait for the result)

Add, replace, upsert, update or delete rows in a table you own, with a safe preview before changes are applied.

Parameters for this action
  • namespacestringrequired
    The namespace (folder) of the table, e.g. 'hr'
  • table_namestringrequired
    The table to change, e.g. 'employee'
  • recordsarray of objectoptional
    Rows to add or merge, for small batches
  • urlstringoptional
    For insert/upsert/overwrite: link to a CSV, Parquet or JSON file
  • file_formatstringoptional
    File type: CSV, Parquet or JSON (detected from the link if left out)
  • delimiterstringoptionaldefault: ","
    CSV separator character (default comma)
  • partition_byarray of stringoptional
    New tables: how to split the data for faster reads, e.g. by month or region (up to 3)
Example prompts
  • >Add these 20 rows to the customers table
  • >Load the orders file from this link into the orders table

Works with every action: wait (False (default): start the job and return a job id to check later. True: wait for the result)

Ask questions of your data lake, export results, and explore its namespaces and tables.

Parameters for this action
  • sqlstringrequired
    The read-only query to run. Your assistant normally writes it for you from your question.
  • limitintegeroptionaldefault: 10
    Maximum number of rows to return (default 10, -1 for all)
  • export_to_s3booleanoptionaldefault: false
    Export the result to a file and return a temporary download link
  • export_formatstringoptionaldefault: "csv"
    Export file format: 'csv' or 'parquet'
  • waitbooleanoptionaldefault: false
    False (default): start the job and return a job id to check later. True: wait for the result
Example prompts
  • >Show total sales by region for last month
  • >List all tables in my data lake
  • >Which customers spent over 500 this year?
  • >Show the 10 most recent orders
  • >Count rows in the orders table per status
  • >Join customers and orders to show each customer's total spend
  • >Show the schema of the sales table
  • >List the namespaces I have

Share a table with another user via an independent grant (column and row restrictions), list your grants, or revoke one.

Parameters for this action
  • namespacestringoptional
    The namespace (folder) of the table to share, e.g. 'hr'
  • table_namestringoptional
    The table to share, e.g. 'employee'
  • share_withstringoptional
    Email address of the person to share with (must have an active account)
  • permissionstringoptionaldefault: "read"
    Read only (default), or write: the person can add, change or delete the rows they are allowed to see, but not change the table itself
  • allowed_columnsarray of stringoptional
    Which columns the person may see (all by default)
  • row_filterstringoptionaldefault: "1=1"
    For 'grant': condition limiting which rows the recipient sees (e.g. "department = 'Engineering'")
Example prompts
  • >Share my sales table with alex@example.com, region and amount only
  • >Share customers with sam@example.com, only EU rows