Integer column not getting mapped in the schema file for BQ

[tanvi-cota]

Describe the bug
I am currently using BigQuery as DB. I have column 'age' set as 'integer' in the table.
But the schema file in cube.js, does not show this column in dimension or in measures.

Expected behavior
All the columns that are present in the table, should be mapped in the dimension.

Version:
0.5.2

[paveltiunov]

@tanvi-cota Hey Tanvi! Thanks for posting this! Do you expect that you'll get something like age in dimensions and averageAge in measures after schema is generated?

[tanvi-cota]

@paveltiunov I would be doing something like 'treatment' in dimensions and age in filter (say > 50) and totalCount in measure

[igorlukanin]

I experienced this bug as well. The use case look very legitimate to me: I was creating the data schema for the tables from the TPC-H benchmark.

For the following table structure, I've got all columns in the data schema except for the int64_field_0 and int64_field_5.

The weird thing is that, indeed, I've got double_field_7 in the schema but mapped to the string type:

int64_field_0  | INTEGER | NULLABLE
string_field_1 | STRING  | NULLABLE
string_field_2 | STRING  | NULLABLE
string_field_3 | STRING  | NULLABLE
string_field_4 | STRING  | NULLABLE
int64_field_5  | INTEGER | NULLABLE
string_field_6 | STRING  | NULLABLE
double_field_7 | FLOAT   | NULLABLE
string_field_8 | STRING  | NULLABLE
string_field_9 | STRING  | NULLABLE

I really think Cube.js should generate dimensions for all columns in the dataset and then let a developer create relevant measures.