cancel
Showing results for 
Search instead for 
Did you mean: 

Google Ads connector - Exclude default tables, exclude CUSTOMER_ID from the very beginning

dmitryungurean
New Contributor II

Hello, dear Fivetran Community.

I have two questions related to the Google Ads connector:

1) Is it possible to include only custom reports in my Google Ads connector from the very beginning? Without all those 76 default tables... I've found a workaround - when it starts to sync, I can go to the 'schema' and uncheck all those tables, but it doesn't really suit my use case.

dmitryungurean_2-1709674375018.png

2) Is there a way not to include the CUSTOMER_ID field in the custom report? I would prefer not to have it in the destination table from the very beginning.

dmitryungurean_1-1709674335184.png

Thank you in advance🙏

Any help will be highly appreciated.

1 ACCEPTED SOLUTION

Hi @SVijay . Hope you are doing well.

Here is the workaround that I ended up with for my use case.

- I retrieve required fields (only those that I need) from Google Ads using Google Ads Scripts

- Put them into a CSV and send it to S3 (using the same Google Ads Script) 

- Collect the CSV files from S3 using Fivetran and push further into the DWH

Not the most elegant solution since I had to spend considerable amount of time learning Google Ads Scripts but it works and overal architecture got simplified considerably.

Million thanks again for your help and insights ... I will use them for API automation a bit later.

View solution in original post

12 REPLIES 12

Dear @dmitryungurean yes, primary keys can't be hashed, as they are needed for identification in the destination and also for various analytical purposes such as joins, unions etc.

I am awaiting our product team's response on the second question. Will let you know ASAP.

Thank you

dmitryungurean
New Contributor II

@SVijay thanks you very much, I will check everything out and come back with feedback.

SVijay
Fivetranner
Fivetranner

Hey @dmitryungurean I have confirmation from our product team.

Re:Google Ads, our sync logic relies on complex key date + customer_id + fivetran_id. Otherwise we might have duplicated records in theory, when different customer_id have the same metrics.

This is in line with what we discussed initially, "CUSTOMER_ID is a primary key for the custom reports and is a required field. This is because Fivetran provides you the capability of setting up multiple Google Ads connectors. One way for you to easily distinguish between similar Google Ads data is via CUSTOMER_ID as everything else w.r.t the other fields can remain same."

You can use your own column blocking / hashing mechanisms within the destination to treat the Fivetran delivered data as "Staging Data" which only you have access to. Once you have the preferred data you need post some massaging of it, you can share that downstream.

As for transformations, you can have a look at Fivetran's dbt packages for Google Ads:

Thank you

@SVijay Thanks a lot for investigating it, greatly appreciate it. I need to think a bit about all the options and consequencies. Will keep you posted.

Always happy to help @dmitryungurean. If you do find a way to load data into BigQuery without the CUSTOMER_ID we'd love to know.

We are always open to receiving feedback and improving.

Thank you.