Merge pull request #4433 from metabase/allow-bigquery-standard-sql

Allow BigQuery standard SQL (inspired by @gitram)

docs/administration-guide/databases/bigquery.md

@@ -7,15 +7,23 @@ Starting in v0.15.0 Metabase provides a driver for connecting to BigQuery direct
 
 1. make sure you have a [Google Cloud Platform](https://cloud.google.com/) account with a Project you would like to use in Metabase.
 * Start by giving this connection a __Name__ and providing your Google Cloud Platform __Project ID__ along with your desired BigQuery __Dataset ID__.  If you don't have a dataset and want to play around with something we recommend copying one of the [sample tables](https://cloud.google.com/bigquery/sample-tables)
-![basicfields](../images/bigquery_basic.png)
+![Basic Fields](../images/bigquery_basic.png)
 * Follow the `Click here` link provided below the __Client ID__ field which will open a new browser tab and guide you through the process of generating OAuth 2.0 credentials for Metabase.  Make sure to choose `Other` for your application type.
-![clientid](../images/bigquery_clientid.png)
+![Client ID](../images/bigquery_clientid.png)
 * take the resulting client ID and client secret and copy them over to Metabase.
-![clientid](../images/bigquery_clientdetails.png)
+![Client Details](../images/bigquery_clientdetails.png)
 * Now follow the link below the __Auth Code__ field for `Click here to get an auth code` which will open a new browser window and authorize your credentials for a BigQuery access token to use the api.  Simply click the `Allow` button.
-![clientid](../images/bigquery_authcode.png)
+![Generating an Auth Code](../images/bigquery_authcode.png)
 * Copy the resulting code provided into the __Auth Code__ field in Metabase.
-![clientid](../images/bigquery_copycode.png)
+![Copying the Auth Code](../images/bigquery_copycode.png)
 * Click the `Save` button!
 
 Metabase will now begin inspecting your BigQuery Dataset and finding any tables and fields to build up a sense for the schema.  Give it a little bit of time to do its work and then you're all set to start querying.
+
+
+## Using Standard SQL
+
+By default, Metabase tells BigQuery to interpret queries as [Legacy SQL](https://cloud.google.com/bigquery/docs/reference/legacy-sql). If you prefer using
+[Standard SQL](https://cloud.google.com/bigquery/docs/reference/standard-sql/) instead, you can tell Metabase to do so by including a `#standardSQL` directive at the beginning of your query:
+
+![Enabling Standard SQL](../images/bigquery_standard_sql.png)

src/metabase/driver/bigquery.clj

@@ -121,6 +121,8 @@
    {:pre [client (seq project-id) (seq query-string)]}
    (let [request (doto (QueryRequest.)
                    (.setTimeoutMs (* query-timeout-seconds 1000))
+                   ;; if the query contains a `#standardSQL` directive then use Standard SQL instead of legacy SQL
+                   (.setUseLegacySql (not (s/includes? query-string "#standardSQL")))
                    (.setQuery query-string))]
      (google/execute (.query (.jobs client) project-id request)))))