Keep samples based on prefered attribute value

aileenzhou

I have a dataset, there are some duplicated DOI. I must keep one of the duplicated DOIs based on 'source' attribute with preference: B>C>A, and delete rest.

For example, the data below, I want to keep row 1261 and 643, delete the rest.
Row     DOI                 Source
18        10.1002/67       A
1261    10.1002/67       B
1400    10.1002/67       C
... ...
643      10.102/et.67    C 
1428    10.102/et.67    A 

Thank you in advance.

lionelderkrikor

Hi @aileenzhou,

In this case, (B>C>A) :  

Then use the same method as in the other thread, but by generating a new attribute called "Source_2" as described  : 

 - Reorder attributes (1/ Source_2 , 2/ DOI)
 - Generate a new attribute (for example called "Source_2") and replace in this new attribute : 
     *B by 1
     *C by 1
     *A by 2
 - Generate concatenation of "Source_2" and "DOI" attributes (via Generate Aggregation attribute)
 - Sort alphabetically the concatenated attributes (via Sort attribute / sorting direction = increasing)
 - Remove duplicates of this concatenated attribute.
 - Split back the concatenated attribute to retrieve the original attributes (without the duplicates) or remove them. 

Take a look at the attached process and tell me if it answer to your need ...

Regards,

Lionel

MartinLiebig

Since Remove Duplicate always keeps the first you can I think sort and then use remove duplicates on the DOI.

Best,

Martin