BigQuery - Delta Tracking using DML MERGE

[@RuxinZhang-jv1oy]

Thank you for the amazing video. Very helpful info.

[@dileepreddy7469]

It's really informative and the way you explained is awesome. Keep up the good work.

[@vinci_irl]

it doesn't work properly, i think the delta_hash column has to be unique or something

[@Philipe_William]

thx for help

[@mayuraggarwal8858]

Hi , Thank you for the awesome video but I think I didn't understand the use case properly. If we have last updated column, then all I need to do is to get records changed after last timestamp like you explained in the end. Right??

[@practicalgcp2780]

Timestamps are often not reliable, the hash comparisons are checking the real data differences. However if you can rely on the last changed time stamp (so they are not timestamp for ingestion but real change from the source data) then it’s fine not to use hash comparisons

[@mayuraggarwal8858]

@@practicalgcp2780 Thank you. It makes sense now. Again thanks a lot for the tutorials.

[@practicalgcp2780]

@@mayuraggarwal8858 no worries, glad you found it useful ;)

[@TheVamos777]

Nice. I wonder where you got that idea 🤔 Glad to see it's going to get seen by a wider audience 😄 Keep the videos coming 👏

[@practicalgcp2780]

Thanks Phil! I can’t remember exactly where the idea of using MERGE for delta tracking came from (I think it was a colleague mentioned something similar in another system and I noticed MERGE can do the same thing in BQ). But the TO_JSON_STRING() idea to avoid null issues were from you😁 don’t you remember?