Question 1

How do sheets and tabs map to tables?

Accepted Answer

One spreadsheet is selected per connection, and each grid tab within it becomes a separate source table, with the output table name derived from the tab name. That maps cleanly onto how people actually use spreadsheets, where one file holds several related but distinct things. If you keep unrelated datasets in one workbook they will still arrive as separate tables rather than being merged.

Question 2

Are reads incremental?

Accepted Answer

No. A spreadsheet does not maintain a change log, so there is nothing to read incrementally from and every run reads the sheet in full. Rows are fetched in batches, and reading stops once a long run of empty rows is reached, which is what keeps a sheet with a few hundred used rows in a large grid from being read to its limits. In practice this is fine, because the sheets people sync are small by data warehouse standards, so a full read on every ETL run is cheap enough to schedule frequently.

Question 3

What happens if someone inserts or reorders a column?

Accepted Answer

This is the failure mode to plan for, because it is silent. Internally fields are positional, identified as column A, column B and so on rather than by their header text. Inserting a column or reordering columns therefore changes what an existing mapping reads, and the sync will happily keep running while writing the wrong values into the right column names. After any structural edit to the sheet, refresh the schema and review the mappings before the next run.

Question 4

How are column types decided?

Accepted Answer

Conservatively. Values default to strings unless the column carries a data validation rule that identifies it as a number, date or boolean, in which case that type is used. If you want reliable typing in the data warehouse, adding validation rules on the sheet is the way to get it. Duplicate identical rows are also deduplicated within a read, which is worth knowing if your sheet legitimately contains repeated rows with no distinguishing column.

Question 5

Why load a spreadsheet into a data warehouse?

Accepted Answer

Because spreadsheets hold the data no system owns. Targets, account tiering, manual corrections, mappings between one system's codes and another's. That data is real and reports depend on it, but it lives outside every pipeline. ETL into a data warehouse puts it next to the system data so joins are ordinary SQL, and it gives you a governed copy with history rather than a link to a file anyone can restructure.

Sync Google Sheets to Snowflake

Sync Google Sheets to Snowflake with Polytomic

Each tab its own table

Headers become column names

Complete snapshot each run

Or sync Snowflake to Google Sheets

Trusted by businesses worldwide

Step-by-step integration guide

Connect Google Sheets in Polytomic

Connect Snowflake in Polytomic

Create a sync between Google Sheets and Snowflake

Map fields and enable the sync

Frequently asked questions

The sync and ETL platform that does it all

Powerful data sync

Advanced analytics

Enterprise security

Easy setup

Ready to connect Google Sheets with Snowflake?