Cast all columns in bigquery to float/string

Question

I'm trying to download a BQ table using python like this: from google.cloud import bigquery

client = bigquery.Client()
SQL_QUERY = """
SELECT *
FROM TABLE
"""
df = client.query(SQL_QUERY).to_dataframe()

I get the following error in the traceback. Looks like, the google-cloud-sdk first converts the table to a JSON format and them dumps to a CSV.

    converter = _CELLDATA_FROM_JSON[field.field_type]
KeyError: 'NUMERIC'

I've two questions:

1. Is there a way to cast all columns in a BQ to float/string before downloading it using Python?
2. How do I identify which columns are causing the trouble?

Answer 1

If you distinctively mention your column names, your queries can become more readable.

Lets say your table has 3 columns ( a , b and c ). If you want to cast these as floats, you can simply modify your query to:

SQL_QUERY = """
  SELECT 
    cast(a as float64) as a, 
    cast(b as float64) as b, 
    cast(c as float64) as c
  FROM TABLE
"""