mcpdata Request an invite

Answers / Get data into your AI

How do I upload a pandas DataFrame so Claude can query it?

Recipe verified on Ran on a private test instance
Short answer

Write it with the cdb client: s.write_table("name", df). It becomes a table Claude can query over MCP straight away, and replace=True swaps in a new version later while keeping the old one.

Steps

  1. Connect your assistant to your instance (this writes, so it needs your own instance with a ReadWrite key from the accounts portal; the public demo is read-only). In Claude Code: 
    claude mcp add --transport http mcpdata https://[you].mcpdata.org/mcp --header "Authorization: Bearer $CDB_API_KEY"

    Replace [you] with your instance's name. Other apps: Connect your AI.

  2. Install the client. Then sign this computer in once (cdb-localdata login --to https://[you].mcpdata.org), or set CDB_API_KEY. 
    pip install https://[you].mcpdata.org/help/download/cdb-0.3.4-py3-none-any.whl
  3. Write the DataFrame.
    import cdb
import pandas as pd

df = pd.DataFrame({"region": ["EMEA", "Americas", "APAC", "EMEA"],
                   "product": ["A", "B", "A", "C"],
                   "sales": [1200.0, 950.5, 430.25, 610.0]})
s = cdb.Session("https://[you].mcpdata.org")
s.write_table("py.sales", df, replace=True)
print(s.execute_sql("SELECT COUNT(*) AS n FROM py.sales").to_dataframe())
  4. Ask your question. Give your AI this prompt:
    Using mcpdata, total sales by region from py.sales.
What your AI does: the tool call it makes for this prompt, and what came back
query(table_name="py.sales", group_by=["region"], agg=[{"func": "sum", "column": "sales", "alias": "sales"}], sort=[{"column": "sales", "desc": true}]) → 3 rows
What you should get · the recipe's real result from its last run
regionsales
EMEA1,810.00
Americas950.50
APAC430.25
Ask Claude with this prompt Ask ChatGPT with this prompt

The assistant needs mcpdata connected (step 1) to run the query.

Why this works

The DataFrame stays a proper table with its column types, rather than text in a prompt. Python and Claude then read the same data, so a number in your notebook and in Claude's answer come from the same rows.

Related questions