Data Uploads
Upload CSV files directly into your BigQuery warehouse with schema inference, column editing, and update-in-place versioning.
Last reviewed September 15, 2026
Data Uploads lets you load a CSV file directly into the BigQuery dataset of your account as a native table. Vendo infers the schema automatically. You can edit column names and types, and Vendo keeps a full upload history.
Uploaded files appear as sources. They show in the Sources list, the Data Catalog and the data tools of the Agent, the same as any connected source.
To start an upload, go to Data → Integrations → Sources, click Add Data Source and select File Upload.
How it works
- Drop a file: drag and drop a
.csvfile, or select one (up to 5 GB). The file uploads directly to secure storage with byte-level progress. - Analyze: Vendo samples the file. It detects the delimiter and whether the first row is a header. It infers a BigQuery column type for every column.
- Review: edit the table name, rename columns, change column types and select the write mode. Vendo shows sample values for each column.
- Load: Vendo streams the file into a BigQuery load job. Small files finish immediately. Large files show live progress (rows loaded, elapsed time) until the load completes.
When the load finishes, Vendo opens the Sources list. The upload shows there as a File Upload source.
Review options
| Option | Choices | Notes |
|---|---|---|
| Table name | Any valid BigQuery identifier | The name of the table created in your dataset |
| Dataset | The dataset of your account (read-only) | Each account has one dataset. You cannot change it. |
| If table exists | Fail / Append rows / Replace all rows | Fail (default) shows an error if the table already exists. Append rows adds rows and allows new columns. Vendo adds new columns to the table, and required columns can become nullable. Replace all rows overwrites the table. |
| First row is header | On / Off | When off, columns are named column_1, column_2, … |
| Column types | STRING, INTEGER, FLOAT64, BOOL, DATE, TIMESTAMP | You can edit the type of each column before loading |
Updating an upload
Each upload has a detail page with its schema, status, and upload history. From there you can:
- Update data: upload a file again into the same target table. The source always points at the latest version, and Vendo keeps the history.
- Semantic Mappings: click Map a table to map the uploaded table to a data type (Events, User Properties, Group Properties or Ad Data). Then map its columns to the canonical fields of that type, for example
email. A mapped table can join into Identity Merge and Customer 360. See Data Model. - Rename the display name of the source.
When you update data, If table exists has 2 choices: Append rows (the default) and Replace all rows. If the new file has a different schema, Vendo shows the added, removed and changed columns before the load. An append with a different schema can fail, or it can load null values into missing columns.
- Delete the upload. This removes the source. You select to keep the BigQuery table or to delete it. If you select Delete data from warehouse, Vendo permanently removes the table. You cannot undo this.
Limits
| Limit | Value |
|---|---|
| File format | .csv only |
| Max file size | 5 GB |
| Max columns | 1,000 |
| Schema inference sample | First 1 MB / 1,000 rows |
Related
- Custom Integrations: other ways to bring custom data into Vendo
- SQL Editor: query uploaded tables directly