Is there a way around casting large integers as string when querying data from BigQuery through R?

Question

Is there a better method for dealing with large integers than casting them as strings when querying data from BigQuery through R via the API?

Here's an MVE showing the problem with the integer appearing as "NA":

> library(bigrquery)
> 
> bq_str <- "
+ SELECT 
+    206932402797274623 AS big_pk
+   ,SAFE_CAST(206932402797274623 AS string) AS string_pk
+ "
> 
> my_df <- bigrquery::query_exec(query = bq_str, 
+                                project = 'XXXXXXXXXXX', 
+                                use_legacy_sql = FALSE,
+                                bigint = "integer64")
0 bytes processed
Warning message:
In converter[[type]](data_m[i, ]) :
  NAs introduced by coercion to integer range
> head(my_df)
  big_pk          string_pk
1     NA 206932402797274623

Here's the code:

library(bigrquery)

bq_str <- "
SELECT 
   206932402797274623 AS big_pk
  ,SAFE_CAST(206932402797274623 AS string) AS string_pk
"

my_df <- bigrquery::query_exec(query = bq_str, 
                               project = 'XXXXXX', 
                               use_legacy_sql = FALSE,
                               bigint = "integer64")
head(my_df)

I am using version 1.1.1 of bigrquery.

Answer 1

Regarding query_exec, this has been deprecated, try using bq_query.

If you are only looking forward to avoid casting to string I recommend to cast to numeric.

Otherwise, you can also use bq_table_download keep in mind to mapped bigint to "integer64".