Status Update
No update yet.
Comments
bl...@google.com
er...@google.com
What's wrong with adding an explicit column to your table schema to refer to the timestamp, which you can set explicitly when inserting the rows? Would that work?
gp...@bendingspoons.com
> What's wrong with adding an explicit column to your table schema to refer to the timestamp, which you can set explicitly when inserting the rows? Would that work?
My system should be schema agnostic, so I used that feature in Legacy SQL to only process recent data.
Also, the system that is writing the data is legacy and we can't really change the schema.
My system should be schema agnostic, so I used that feature in Legacy SQL to only process recent data.
Also, the system that is writing the data is legacy and we can't really change the schema.
Description
Im' trying to look at the rows that were inserted AFTER a certain timestamp.
The important part is that this feature allows me to be completely agnostic regarding the table structure.
How this might work:
```
SELECT *
FROM my_table
FOR SYSTEM TIME BETWEEN <timestamp1> AND <timestamp2>
```
Or, by removing timestamp2
```
SELECT *
FROM my_table
FOR SYSTEM TIME AFTER <timestamp1>
```
If applicable, reasons why alternative solutions are not sufficient:
This function used to work well in Legacy SQL but the porting in Standard SQL left behind this functionality which I find very useful. Now we're migrating to Standard SQL and loving the new functionalities, but in order to make this part work we're missing this critical feature.
Other information (workarounds you have tried, documentation consulted, etc):
I tried Legacy SQL, but I'm having problems with manipulation of nested fields, which is handled very gracefully in Standard SQL, but without this feature I can't be agnostic from the tables.