Queries from the Node.js BigQuery client v2.0.0 and higher are not returning results

[d-torres]

Running a query in a Node.js 8 environment using the BigQuery Client Library, v3.0.0 results in an empty response returned to the caller.

Reproduction steps and observed behavior:

Sample query: "with foo as (select 'Travis' as name) select count(*) as total from foo"

Expected behavior: The query() method should return COUNT = 1 after running the aggregate query from above.

Actual behavior: It's currently returning empty for bq client 2.0.0 and above.

[sduskis]

@callmehiphop, can you please take a look at this?

[callmehiphop]

@d-torres I just tested this against master and got the following result

[
  {
    "total": 1
  }
]

[stephenplusplus]

I'm not sure how, but the exact version of Node 8 might matter, since we only support 8.10 and up:

nodejs-bigquery/package.json

Line 8 in d64d892

"node": ">=8.10.0"

.

[JustinBeckwith]

The only reason that was set to 8.10 was because of HTTP/2 in the APIs that used grpc-js. Shouldn't matter here I don't think, but good to know regardless.

[cmath]

Same thing happened to us. All of a sudden Select queries stopped working when running from cloud functions. Inserts are fine. When testing locally everything works as it should...

[d-torres]

@callmehiphop - Could we take a look at this from the context of a cloud function to see if we're able to reproduce?

[callmehiphop]

Confirmed not to be working on GCF, using the query method I started hitting a Maximum call stack size exceeded error when requesting a single page of results. For a temporary workaround users can create a query job like so

bigquery.createQueryJob(`SELECT * FROM foo`, (err, job) => {
  if (err) // error handling omitted
  job.getQueryResults((err, rows) => {
    if (err) // error handling omitted
  });
});

Edit: Here's the error stack trace

RangeError: Maximum call stack size exceeded
    at replaceProjectIdToken (/srv/node_modules/@google-cloud/projectify/build/src/index.js:29:31)
    at replaceProjectIdToken (/srv/node_modules/@google-cloud/projectify/build/src/index.js:40:30)
    at value.map.v (/srv/node_modules/@google-cloud/projectify/build/src/index.js:31:32)
    at Array.map (<anonymous>)
    at replaceProjectIdToken (/srv/node_modules/@google-cloud/projectify/build/src/index.js:31:23)
    at replaceProjectIdToken (/srv/node_modules/@google-cloud/projectify/build/src/index.js:40:30)
    at replaceProjectIdToken (/srv/node_modules/@google-cloud/projectify/build/src/index.js:40:30)
    at value.map.v (/srv/node_modules/@google-cloud/projectify/build/src/index.js:31:32)
    at Array.map (<anonymous>)
    at replaceProjectIdToken (/srv/node_modules/@google-cloud/projectify/build/src/index.js:31:23)%  

[callmehiphop]

Looks like the culprit is this line, it causes projectify to be called recursively quite a bit. For a fix I want to suggest not caching a job instance, but rather the job ID which we can use to grab a new job instance.

[melissammbeff]

I'm using @google-cloud/bigquery": "^3.0.0" and I'm having exactly this problem calling createQueryStream function. What should I do?

[callmehiphop]

@melissammbeff are you able to upgrade to the latest version (4.4.0)?

[melissammbeff]

Hi @callmehiphop, so, unfortunately not so far. I understood that this problem was solved on 2.0.6, so, why I'm still having the same problem on 3.0.0?

[callmehiphop]

@melissammbeff I'm not sure where 2.0.6 came from, but the PR that resolved this issue (#465) was not released until 4.1.1.