将 Google Data Studio 社区连接器与 BigQuery 结合使用时的时间戳查询问题

Question

I'm having an issue where Google Data Studio is sending bad timestamp data to my community connector, so that when I try and filter on a date, I get an error. I have two BigQuery TIMESTAMP type fields (named timestamp and created_at ), both being passed through to my community connector without modification. Once I add a date filter to the reports (in order to do time series, or regular filtering), like so:

then my queries from the connector (viewed in my BigQuery project history) begin to fail like this:

Could not cast literal "20200825" to type TIMESTAMP at [1:677]

The query in BigQuery looks something like this:

SELECT t0.created_at, SUM(t0.sum_metric) AS t0_qt_1z4br3iwbc FROM (SELECT field1, field2, field3) from data_name.table_name where user_identifier in (2)) AS t0 WHERE (t0.created_at >= '20200825' AND t0.created_at <= '20200831') GROUP BY t0.created_at ORDER BY t0.created_at ASC;

This really feels like a bug with the community connector regarding BigQuery. Is there some way around this? Am I just doing something wrong I'm not seeing?

Answer 1

Ok, I solved this. In your community connector you'll need to add this to your config:

config.setDateRangeRequired(true);

this will send a startDate and endDate parameter with your getData request (it defaults to 28 days). Access them in getData() like so:

var startDate = request.dateRange.startDate;
var endDate = request.dateRange.endDate

and then use them in your query as necessary.

As a side note, if you're storing a timestamp field in Google Data Studio and making a community connector, you'll need to set up a calculated field so that the reports treat it appropriately (display a human readable date instead of a timestamp). I'm first reading the timestamp field out of the database as a string like: STRING(timestamp, 'UTC') AS timestamp and then using this value to create a dimension. This is done in the schema like the following (use parsing appropriate to your field if it's different):

fields.newDimension()
  .setId('date_timestamp')
  .setName('date_timestamp')
  .setDescription('Timestamp as a date.')
  .setFormula("TODATE($timestamp, '%Y-%m-%d %H:%M:%S%z', '%Y%m%d')")
  .setType(types.YEAR_MONTH_DAY)