I have a simple load step
DEFINE JOB Run_Load
APPLY $INSERT TO OPERATOR
TargetTable = @TargetDatabase || '.' || @TargetTable,
LogTable = @TargetDatabase || '.' || @TargetTable || '_LOG',
ErrorTable1 = @TargetDatabase || '.' || @TargetTable || '_E1',
ErrorTable2 = @TargetDatabase || '.' || @TargetTable || '_E2'
SELECT * FROM OPERATOR($FILE_READER( DELIMITED @TargetDatabase || '.' || @TargetTable ) );
If I have understood correctly, the schema will be inferred from the target table
This is working fine when I have a table defined as:
Please explain what you mean by:
"This approach seems to have worked ok with dates previously."
By "previously", do you mean TPT used to work and now does not?
BTW, when we say we can infer the schema from a table, this means we take the column data types.
We do not look at the format.
That is resolved by the DBS.
Why did you define the column as TIMESTAMP(6) and not TIMESTAMP(0)?
Your data has no fractional seconds?
Kindly refer the below link for TPT script details: https://teradataworld.wordpress.com/teradata-paral
The data is moved by a date datatype , the format is just for representation., What i can understand is your data is in YYYYmm/dd and you need to load data in dd/mm/yyyy format, in such scenario you need to export and import , once data is loaded in the table it will be represented in the format you have specified.
Hi Steve - by 'previously' I meant I thought it was working this way; obviously it never did, but as you say the DBS was being helpful by handling the date formats I sent anyway. I understand now, thank you.
I've reverted to having to generate the schema of the producer file myself, so I can specfy the date format like:
VARDATE(20) FORMATIN ('ddMMMyyyy') FORMATOUT ('YYYY-MM-DD')
Thanks for advice