Delta Live Tables with CDC and Database Views with Lower Case Names

zesdatascience — Tue, 03 May 2022 23:48:09 GMT

Hi,

I am testing out creating some Delta Live Tables using Change Data Capture and having an issue where the resulting views that are created have lower case column names. Here is my function I am using to ingest data:

def raw_to_ods_merge(table_name,source_stream,file_description,

ods_path,primary_keys,sequence,table_prop,column_comments):

ingest_name = table_name + '_ingest'

output = source_stream \

.withColumn('ETL_InputFile',F.input_file_name()) \

.withColumn('ETL_LoadDate',F.lit(datetime.datetime.now()))

@dlt.view(

name = ingest_name

)

def source_ingest():

return (output)

dlt.create_target_table(name = table_name,

comment = file_description,

path = ods_path,

table_properties=table_prop,

schema = output.schema

)

dlt.apply_changes(

target = table_name,

source = ingest_name,

keys = primary_keys,

sequence_by = sequence

)

This is resulting in an __apply_change... table name with column names in mixed case as expected. However the view that gets created sets the output column names to lower case. e.g.

CREATE VIEW `raw`.`shopper_panel_korea` (

`duration` COMMENT 'Length of time for which the metrics apply.',

`period` COMMENT 'Date for which the metrics apply, the end of the date range.',

`product`,

`spend_1000000_krw`,

`product_percent_of_category_value`,

`retailer_percent_of_value`,

`volume_1000_kg`,

`product_percent_of_category_volume_1kg`,

`retailer_percent_of_volume_1kg`,

`buyers_1000`,

`penetration_percent`,

`frequency`,

`trips_1000`,

`performance`,

`channel`,

`domestic_or_import`,

`etl_inputfile`,

`etl_loaddate`)

TBLPROPERTIES (

'transient_lastDdlTime' = '1651620122')

AS SELECT `Duration`,`Period`,`Product`,`Spend_1000000_KRW`,`Product_Percent_of_Category_Value`,`Retailer_Percent_of_Value`,`Volume_1000_kg`,`Product_Percent_of_Category_Volume_1kg`,`Retailer_Percent_of_Volume_1kg`,`Buyers_1000`,`Penetration_Percent`,`Frequency`,`Trips_1000`,`Performance`,`Channel`,`Domestic_or_Import`,`ETL_InputFile`,`ETL_LoadDate` FROM `raw`.`__apply_changes_storage_shopper_panel_korea` WHERE __DeleteVersion IS NULL

Is there something I am doing wrong or is this an issue?

Thanks in advance for your help,

Stu

Re: Delta Live Tables with CDC and Database Views with Lower Case Names

zesdatascience — Thu, 19 May 2022 04:45:49 GMT

Hi @Kaniz Fatma

Thanks for your email. Yes, I did look through the article shared. I am set up with Delta Live Tables and pipelines are working OK generally.

It was really just the one issue mentioned in the post with the lower case column names in the views when view through the Hive catalog. It was wondering whether this is a defect with the view creation which is handled automatically through Delta Live tables? Not sure other than this, where I would log that though?

One thing I did notice was it looks fine through Databricks SQL. Unless it is something to do with the runtime version I was using, I'll check I am running on the latest and see if the issue is still there.

Thanks for your help,

Stu

Re: Delta Live Tables with CDC and Database Views with Lower Case Names

zesdatascience — Tue, 07 Jun 2022 22:09:57 GMT

Hi @Kaniz Fatma

Not found a solution just yet, but not a priority as most users will be accessing through Databricks SQL, so no further assistance required right now.

Thanks

topic Re: Delta Live Tables with CDC and Database Views with Lower Case Names in Data Engineering

Delta Live Tables with CDC and Database Views with Lower Case Names

Re: Delta Live Tables with CDC and Database Views with Lower Case Names

Re: Delta Live Tables with CDC and Database Views with Lower Case Names