Skip to main content
Version: edge

gbq

note

Authentication happens over the GCP autentication

gbq_writer

The gbq_writer makes it possible to write events to Google BigQuery by using its gRPC based [storage API v1].

Configuration

optiondescription
table_idThe identifier of the table in the format: projects/{project-name}/datasets/{dataset-name}/tables/{table-name}
connect_timeoutThe timeout in nanoseconds for connecting to the Google API
request_timeoutThe timeout in nanoseconds for each request to the Google API. A timeout hit will fail the event.
request_size_limitSize limit (in bytes) for a single AppendRowsRequest. Defaults to the quota documented by Google (10MB)
tokenThe authentication token see GCP autentication

The timeouts are in nanoseconds.

use std::time::nanos;

define connector gbq from gbq_writer
with
config = {
"table_id": "projects/tremor/datasets/test/tables/streaming_test",
"connect_timeout": nanos::from_seconds(10),
"request_timeout: nanos::from_seconds(10)
"token": "env", # required - The GCP token to use for authentication, see [GCP authentication](./index.md#GCP)
}
end;

Metadata

The $gbq_writer.table_id field can be set, to send an event to a table other than the one globally configured. A new writestream will be opened per distinct table_id.

Example of setting the table_id metadata in a script:

let $gbq_writer = {
"table_id": "projects/my_project/datasets/my_dataset/tables/my_table"
};

### Payload structure

The event payload sent to the `gbq_writer` connector needs to be a [`record`](../../language/expressions.md#records) with field names being the table column names
and the values need to correspond to the table schema values.

Tremor values are mapped to Google Bigquery schema types according to the following value mapping. You need to provide the tremor value type on the right to feed a column of the left side.

| Google Bigquery type (gRPC type) | Tremor Value | Format | Examples |
|------------------------------------|----------------|-------------------------------------------------------------|---------------------------------------------------------|
| `Numeric` | `string` | `"X.Y"` (no thousands separator, `.` as decimal point) | `"0.123"`, `"123.0"`, `"1.234"` |
| `Bignumeric` | `string` | `"X.Y"` (no thousands separator, `.` as decimal point) | `"0.123"`, `"123.0"`, `"1.234"` |
| `Int64` | `integer` | | `1234` |
| `Double` | `float` | | `1.234` |
| `Bool` | `bool` | | `true`, `false` |
| `Bytes` | `binary` | | |
| `String` | `string` | | `""`, `"badger" |
| `Date` | `string` | `"YYYY-[M]M-[D]D"` | `"2015-1-14"`, `2022-09-13` |
| `Time` | `string` | `"[H]H:[M]M:[S]S[.DDDDDD|.F]"` | `"0:1:2"`, `"00:01:02"`, `"00:00:00.000123"` |
| `Datetime` | `string` | `"YYYY-[M]M-[D]D[( |T)[H]H:[M]M:[S]S[.F]]"` | `"2015-06-13T00:01:02"`, `"2015-06-13T00:01:02.000001"` |
| `Geography` | `string` | [OGC Simple Features](https://www.ogc.org/standards/sfa) | |
| `Interval` | `string` | `"[sign]Y-M [sign]D [sign]H:M:S[.F]"` | `+10-0 +D00:01:02` |
| `Timestamp` | `string` | `"YYYY-[M]M-[D]D[( |T)[H]H:[M]M:[S]S[.F]][time zone]"` | `"2015-06-13T00:01:02.000001Z"` |
| `Struct` | `record` | | `{"a": 123, "b": "c"}` |



#### Example

For a table defined like:

```bigquery
CREATE TABLE test (
id INT64,
payload STRUCT<a INT64, b INT64>,
name STRING
)

An example event payload would be:

{
"id": 1234,
"payload": {"a": 1, "b": 2},
"name": "Tremor"
}