How to store numeric type in BigQuery using node js client for streaming insert?

Question

Given that numeric type in BigQuery supports limits way beyond the number type of Javascript, how do I store a numeric type using the node js client Table.insert(rows, options, callback) method? For handling big numbers using string type, I was considering using https://github.com/MikeMcl/big.js. If I send a string value (containing a valid number beyond javascript's number limit) for a numeric field in my BigQuery table, will it work?

Answer 1

You need to treat the numbers that are too big for JavaScript as strings.

For example, JS doesn't know how to handle -99999999999999999999999999999.999999999:

Meanwhile this code works in NodeJS to insert that huge number into BigQuery:

BigQuery = require('@google-cloud/bigquery');
bigquery = BigQuery({projectId: 'my-project'});
dataset = bigquery.dataset('tt');
table = dataset.table('testfun_numeric');

rows = [{
  name:"whatever",
  numi: "-99999999999999999999999999999.999999999"
}];
table.insert(rows);

And in BigQuery:

SELECT name, numi 
  , numi=numi+1000 better_comparison
FROM `fh-bigquery.tt.testfun_numeric` 

I previously created the table in BigQuery with

CREATE TABLE `tt.testfun_numeric`
AS
SELECT '' name, CAST(0 AS NUMERIC) numi 
LIMIT 0

FWIW, I also tried BigInt(Number.MAX_SAFE_INTEGER) in NodeJS, but the library wouldn't send it correctly to BigQuery - unless I converted it into a string first.