GOOGLE DRIVE → SNOWFLAKE
Sync CSV, Excel, TSV, or Google Sheets from Drive into Snowflake in under 10 minutes — automated schema discovery and incremental sync.
Supaflow's Google Drive connector handles the file-type quirks that break naive pipelines for bulk file drops: Excel formula cells, duplicate CSV headers, TSV mime-type variance, and mixed-schema folders. For live spreadsheet workflows where worksheets are the table unit, the companion Google Sheets to Snowflake page goes deeper. Every connector is included on every Supaflow plan — you pay only for the compute your pipelines consume.
Used by finance teams consolidating vendor CSV drops for month-end close, and by marketing analytics teams pulling ad-platform exports into Snowflake — with no per-row fees.
Every row below is an actual capability in the Google Drive connector, not a forward-looking promise.
| Feature | How it works | Limit / caveat |
|---|---|---|
| Supported file types | CSV, TSV, Excel (.xls and .xlsx), or native Google Sheets — each with format-specific parsing and type inference. | One file type per source. To sync multiple file types from the same Drive, configure one source per file type (each gets its own pipeline and tables). |
| Authentication | Two auth modes: user OAuth 2.0 (Google sign-in flow, refreshable tokens, fastest setup) or a Google service account (granted Viewer access on the target Drive folders, best for unattended pipelines). Pick per source. | Token refresh happens automatically to prevent mid-sync failures, regardless of mode. |
| Incremental sync | Only files updated since the last successful run are re-read. Tracking happens per file, not per folder. | Renamed or moved files keep their Drive file ID, so they continue syncing without re-config. |
| Schema discovery | Supaflow samples a configurable number of files per folder, unions headers from the sampled subset with first-seen casing preserved, and infers types across the aggregated sample. The discovered schema is then used to read the entire folder. | Columns that exist only in unsampled files are not in the discovered schema until schema is rediscovered. Increase the schema sample size for high-variance folders, or rediscover after large batches of new files arrive. |
| Rate limiting | Google rate limits and transient listing or read failures are retried automatically before a file is marked failed. | Folder listing page size is configurable for very large Drive folders. |
| Cancellation safety | Cancelled jobs stop promptly between folder listing, file reads, and parsing checkpoints. | A cancelled job stops without continuing to consume Google API quota in the background. |
| Audit metadata | Every row loaded into Snowflake carries source file metadata for lineage and debugging. | — |
| Destination | Lands directly in Snowflake with generated tables, Snowflake-ready types, and incremental merges where applicable. | — |
Every Supaflow connector is included on every plan at no extra cost. You pay only for compute consumed, measured in Supaflow Credits (1 credit = 1 compute-hour on a Small node). No per-row fees.
Excel formula cells break naive parsers when external workbook references are missing. CSV files with duplicate headers silently collapse columns in most readers. TSV files come back as `text/plain` half the time depending on which BI tool exported them. Our connector handles each specifically.
A Drive folder rarely holds files with identical schemas. Supaflow samples a configurable number of files per folder, unions their headers while preserving first-seen casing and insertion order, and infers types across the aggregated sample. No "first file defines schema" cliff. For folders where schema variance is wide, increase the sample size or rediscover schema after large batches of new files arrive.
Every one of these is something our connector handles specifically. A generic pipeline built in Airflow or a basic ELT tool will hit at least three of them in production.
Failure mode: Excel formula cells can evaluate to #REF!, #N/A, or simply fail POI evaluation when referenced workbooks are not present. Naive readers throw and abort.
Evidence: Tested across workbooks with external references, circular dependencies, and deprecated functions from older Excel versions.
Fix: Three-tier fallback: numeric evaluation first, cached string value second, null on both failures. Pipeline continues; bad cells become nulls in Snowflake.
Failure mode: Real-world CSV files routinely contain duplicate column names ("amount", "amount", "amount") from hand-maintained spreadsheets. Naive parsers either error or silently collapse columns.
Evidence: Encountered in customer finance exports where amount appears for gross, tax, and net in the same header row.
Fix: Append positional suffixes: amount, amount_2, amount_3. All columns land in Snowflake; analysts see the deterministic suffix pattern.
Failure mode: A single Drive folder often holds files with overlapping but non-identical schemas — one CSV has a status column the others do not. A strict "first file defines schema" approach loses columns.
Evidence: Standard pattern in data-ops folders where different teams drop files with slightly different export tooling.
Fix: Supaflow unions headers across sampled files while preserving first-seen casing and insertion order. The discovered schema covers every column found in the sample; columns that exist only in unsampled files require a schema rediscovery to land. Missing values within the discovered schema are nulls.
Failure mode: Drive categorizes tab-separated files inconsistently — some come back as text/tab-separated-values, others as text/plain. A single mime filter misses half of them.
Evidence: Confirmed against real Drive folders with TSV exports from BI tools (some set the right mime type, some default to text/plain).
Fix: File discovery matches both mime types. File-extension check gates which are parsed as TSV vs generic text.
Pick an auth mode: sign in with a Google user account via OAuth (fastest setup), or share the target Drive folders with a Google service account (best for unattended pipelines). Each Google Drive source is configured for one file type (CSV, TSV, Excel, or Google Sheets).
Google Drive source docs→Select the folder containing the files for this source. Supaflow automatically discovers files of the configured file type, samples them to infer schema, and creates one table per worksheet (Excel/Sheets) or per folder (CSV/TSV).
Connect your Snowflake warehouse. Supaflow creates Snowflake-ready tables and preserves source keys where relevant.
Snowflake destination docs→Pick a cron or interval. Incremental syncs pull only files changed since the last successful run.
Schedules docs→Supaflow authenticates with either user OAuth 2.0 or a Google service account, discovers files in the folders you select, samples a configurable number of files to infer schema, and loads the data into Snowflake tables. Incremental syncs pull only changed files. Schema evolution is handled when columns appear in newly sampled files; for folders that gain new columns over time, rediscover schema periodically.
Yes. Native Google Sheets, CSV, TSV, and Excel (.xls and .xlsx) are all supported. Each Google Drive source is configured for one file type, so to sync more than one file type from the same Drive you create a separate source per file type. Each Google Sheets worksheet becomes its own Snowflake table, and worksheets renamed between syncs continue working because Supaflow tracks the stable worksheet identity behind the name.
Fivetran and Hevo often price by rows or volume, which makes spreadsheet ingestion unpredictable — a finance team dropping a large file can spike your bill mid-month. Every Supaflow connector is included on every plan at no extra cost. You pay only for compute consumed, measured in Supaflow Credits (1 credit = 1 compute-hour on a Small node). No per-row fees. See the pricing page for the credit packages and free tier.
Yes. On every sync, Supaflow re-reads only files changed since the last successful run. Renamed or moved files keep their Google Drive file identity, so they continue syncing without reconfiguration.
Google's API rejects export-to-xlsx for sheets over 10 MB. Supaflow catches this specific error, logs the file name, and continues the sync without failing the pipeline. The operator gets a clear signal that the sheet needs to be split or exported manually. Most teams only hit this for a handful of sheets — the rest of the sync completes normally.
Yes. Supaflow's sync agent can run inside your own VPC so that Drive credentials, file content, and Snowflake credentials never leave your network. The control plane is managed; the data plane can be managed or self-hosted. Same connector, same features.
Duplicate headers are resolved with positional suffixes: "amount", "amount_2", "amount_3". Every column lands in Snowflake deterministically, so analysts can distinguish them without rewriting the source file.
The Google Drive source and Snowflake destination, plus other Supaflow connectors you can pair into a Snowflake pipeline.
Source connector overview and capabilities.
Destination connector overview and capabilities.
Sync Salesforce objects into Snowflake.
Companion page for live spreadsheet workflows (Sheets-as-table).
Replicate Postgres tables into Snowflake.
Browse the full catalog of sources and destinations.
Full setup guide for both OAuth and service account auth, permissions, and troubleshooting.
Connect your Snowflake warehouse, role requirements, type mapping, and sync semantics.
Long-form walkthrough from zero to production, with screenshots.
Every connector is included on every plan. Pay only for compute consumed (Supaflow Credits).
Every connector is included on every Supaflow plan — you pay only for the compute your pipelines consume. Incremental sync, schema evolution, set up in under 10 minutes.