Snowflake 中 COPY INTO 命令中的正则表达式

Question

I have few CSV files in azure blob storage, and we are using COPY INTO command to load the files in snowflake table. The problem is: The file system is: container >> folder (Ex: account) >> Number of files like 2011-09.csv 2011-10.csv likewise and account folder also has a sub-folder 'Snapshot' which also has files that has similar data but with different name like 2019-11_1654478715.csv So while using COPY INTO command, the target table in Snowflake is populated with duplicate rows.

Iam using this one:

copy into BINGO_DWH_DEV.LANDING.CRM_ACCOUNT_TEMP from 'azure://abc.blob.core.windows.net/abc-abc/account' credentials=(azure_sas_token= 'abc') ON_ERROR='CONTINUE' FILE_FORMAT=(type=csv field_delimiter=',' FIELD_OPTIONALLY_ENCLOSED_BY='"');

Any ideas where I can use COPY INTO command with regular expression that can pick only the files like '2011-09.csv' and not the files from the Snapshot folder.

Appreciate your help

Answer 1

You can use pattern keyword as regular expressions to insert files based on pattern.

Please refer to the Snowflake documentation .

Example:

copy into emp_basic
  from @%emp_basic
  file_format = (type = csv field_optionally_enclosed_by='"')
  pattern = '.*2011-19.*.csv.gz'
  on_error = 'continue';

It depends on how you set the stage location (Azure blob or S3 or GCP). Let's say that your files get landed in the "folder" s3://yourbucket/folder1/[filename],gz . And that you've set your stage to point to s3://yourbucket used pattern:

pattern='.*2011-09*.csv.*.gz'

Then it will scan all files under s3://yourbucket .

If however your stage has been setup to point to the folder s3://yourbucket/folder1/ and the pattern used is:

pattern='.*2011-09.*csv.*.gz'

Then it will look only in s3://yourbucket/folder1/ .