The Scenario
You joined the data engineering team at a fintech company three weeks ago. On your first day, someone handed you access to Snowflake and said "the schema documentation is in a Confluence page from 2023." The Confluence page has 40 tables listed. Snowflake has 300.
Your manager wants a complete data dictionary by end of week — every database, every schema, every table you have access to, with row counts, byte sizes, owners, and creation dates. It's supposed to be the foundation for a data governance initiative. It's supposed to be comprehensive. And nobody told you that building it manually means running SHOW TABLES and SHOW DATABASES dozens of times and copying results into a sheet one schema at a time.
The bad version:
- Run SHOW DATABASES in SnowSight, copy the output, paste into a sheet
- For each database, run SHOW SCHEMAS, repeat the copy-paste cycle
- For each schema, run SHOW TABLES, copy the output, add it to the accumulating sheet — and realize row 1 doesn't line up with row 2 because the column order changed between schemas
The dictionary isn't done. It's a stack of raw SHOW outputs that still needs to be normalized, sorted, deduplicated, and formatted before it's useful to anyone. You've spent four hours and you have a mess.
The Easy Way: One Prompt in SheetXAI
SheetXAI is an AI agent that lives inside your Google Sheet. It reads the sheet, connects to Snowflake through its built-in integration, and can traverse your entire account hierarchy — databases, schemas, tables — and write a structured inventory into the sheet in one pass.
Open a blank Google Sheet and paste this into the SheetXAI sidebar:
List all Snowflake databases I have access to, then for each database list all schemas, and for each schema list all tables with their row count, size in bytes, owner, and creation date — write everything into this Google Sheet with one row per table and columns: database, schema, table_name, row_count, size_bytes, owner, created_date
What You Get
- One row per table across your entire account
- Columns: database, schema, table_name, row_count, size_bytes, owner, created_date
- Sorted and structured — not a raw SHOW dump
- Any table you don't have SELECT access to is noted rather than silently omitted
What If the Data Is Not Quite Ready
You only need tables above a certain size threshold
List all Snowflake databases I have access to, then for each database and schema, find all tables with row_count greater than 10,000 — write database, schema, table_name, row_count, and size_bytes into this sheet sorted by row_count descending
You need to limit the catalog to specific databases
For Snowflake databases ANALYTICS and PROD_DW only, list all schemas and all tables within those schemas, write database, schema, table_name, row_count, size_bytes, owner, and created_date into this sheet with one row per table
You want to flag tables with no recent data activity
List all tables in Snowflake databases ANALYTICS and PROD_DW, include created_date and row_count, write into this sheet, then add a column called stale that says YES for any table with row_count = 0 or created_date older than 180 days
Full catalog with a completeness check and size summary at the bottom
List all Snowflake databases and schemas I have access to, for each schema list all tables with row_count, size_bytes, owner, and created_date — write one row per table into this sheet, then add a summary row at the bottom showing total tables counted, total size_bytes across all tables, and count of tables with row_count = 0
The pattern: describing the shape of the output — what gets written, what gets flagged, what gets summarized — is part of the same prompt as the data retrieval.
Try It
Get the 7-day free trial of SheetXAI and open a blank Google Sheet, then ask it to enumerate your entire Snowflake account hierarchy into a structured data dictionary. Also see export a schema table inventory for a targeted version of this workflow, or return to the Snowflake integration overview.