Rename all column names for joining purpose in big query

问题

I need to join two tables with exactly the same column names. I will need to rename the columns before the joining step. Each table contains 100+ columns. I wonder if there's any way to add prefix or suffix to rename all the columns, rather than manually change them with AS. I'm using standard SQL on BigQuery. I put an example in below for illustration. I wonder if there's any function in BigQuery like:

UPDATE
  inv
SET
  CONCAT(column_name, '_inv') AS column_name

... example ... thank you in advance!

WITH
  inv AS (
  SELECT
    '001' AS company,
    'abc' AS vendor,
    800.00 AS transaction,
    'inv' AS type
  UNION ALL
  SELECT
    '002' AS company,
    'efg' AS vendor,
    23.4 AS transaction,
    'inv' AS type ),
  prof AS (
  SELECT
    '001' AS company,
    'abc' AS vendor,
    800.00 AS transaction,
    'prof' AS type
  UNION ALL
  SELECT
    '002' AS company,
    'efg' AS vendor,
    23.4 AS transaction,
    'prof' AS type )
SELECT
  inv.*,
  prof.*
FROM
  inv
FULL JOIN
  prof
USING
  (company,
    vendor,
    transaction)

回答1:

Instead of SELECT inv.*, prof.* which obviously ends up with Duplicate column names in the result are not supported. ... use SELECT inv, prof as shown below

#standardSQL
WITH inv AS (
  SELECT'001' AS company,'abc' AS vendor,800.00 AS transaction,'inv' AS type UNION ALL
  SELECT'002' AS company,'efg' AS vendor,23.4 AS transaction,'inv' AS type 
), prof AS (
  SELECT'001' AS company,'abc' AS vendor,800.00 AS transaction,'prof' AS type UNION ALL
  SELECT'002' AS company,'efg' AS vendor,23.4 AS transaction,'prof' AS type 
)
SELECT inv, prof
FROM inv FULL JOIN prof
USING (company, vendor, transaction)

result :

Row inv.company inv.vendor  inv.transaction inv.type    prof.company    prof.vendor prof.transaction    prof.type    
1     001           abc         800.0             inv         001             abc           800.0               prof     
2     002           efg         23.4              inv         002             efg           23.4                prof

As you can see resulted row now has two structs/records - each holding respective entry from respective table

I think this is your best option here from all practical reasons

来源：https://stackoverflow.com/questions/49942891/rename-all-column-names-for-joining-purpose-in-big-query

标签

sql

google-bigquery