Expand my Community achievements bar.

SOLVED

Question with .overwrite partitions

Avatar

Level 6

Hi there,

 

We have a .overwrite partition for customer profiles that is broken into 26 partitions. If our data looks like this:

Customer ID (Cross Device ID)Key Value PairsIn partition
1"product_id" = "123", "activated" = "Y".02
1"product_id" = "223", "activated" = "N".26

 

I am wondering for customer 1, which profile would they get? Would they get product 123, or product 223, or both? I am hoping it's both but since this is a .overwrite method, I am not sure if it will get product 223 because it's in the .26 partition which will get processed after .02 and overwrite it?

 

Thanks,

Alan

1 Accepted Solution

Avatar

Correct answer by
Employee Advisor

Hello @akwankl ,

 

Hope you are keeping well!

In the above example, only product_id=223 and activated=n will stay as we are talking about overwrite files.

AAM will consider only the last record for that customerID when it comes to overwrite files.

 

Regards

Ankita Sodhi

View solution in original post

4 Replies

Avatar

Correct answer by
Employee Advisor

Hello @akwankl ,

 

Hope you are keeping well!

In the above example, only product_id=223 and activated=n will stay as we are talking about overwrite files.

AAM will consider only the last record for that customerID when it comes to overwrite files.

 

Regards

Ankita Sodhi

Avatar

Level 6

Hi @Ankita_Sodhi 

 

Just wanted to follow up on this, even if the files are partitioned files being sent to AAM, it will still only consider the last record?

Avatar

Employee Advisor

Hello @akwankl ,

 

In case of overwrite files, yes the last record will be considered and will overwrite all others.

Avatar

Level 6

Hi @Ankita_Sodhi,

 

Bringing this thread backup as we have been running into few issues.

 

Couple questions:

1) In my original example, how can we structure our files to qualify customer ID1 to both product_id = "123" & product_id = "223" traits?

2) Current state, our files are actually structured per my listed example but the files are spread across 600 partitions. We see some customers actually qualify for multiple traits (aka multiple product_ids, ie. product_id = 123, product_id = 223), but some customers on the contrary qualify for only one product when they should be in multiple.

 

The fact that some customers are qualifying for multiple products contradicts with what you said, so I am not sure what's the right answer.

 

Thanks!