Hi all,
since I'm new to this, I apologize if I'm flogging a dead horse, and RTFMs with pointers are welcome.
That being said, I have an input file like this:
group1,alice
group1,bob
group2,peter
group2,paul
group2,mary
This I want to transform to smth. along the lines of:
"group1","alice;bob"
"group2","peter;paul;mary"
Can this be done with clover ETL? I can't seem to think of a way.
Normal reformats work on individual records without knowledge about previous or next records.
I could make a reformat do nothing (just collect data and save the last seen group name) for records until the first element changes, then write an output record ... but I would prolly be missing the last record.
Was trying to work with lookup tables, but I don't have unique keys.
Would it be feasible to use transforms purely for building up internal hashes and writing output from finalize?
Any other ideas, thoughts?
TIA,
Karl.
since I'm new to this, I apologize if I'm flogging a dead horse, and RTFMs with pointers are welcome.
That being said, I have an input file like this:
group1,alice
group1,bob
group2,peter
group2,paul
group2,mary
This I want to transform to smth. along the lines of:
"group1","alice;bob"
"group2","peter;paul;mary"
Can this be done with clover ETL? I can't seem to think of a way.
Normal reformats work on individual records without knowledge about previous or next records.
I could make a reformat do nothing (just collect data and save the last seen group name) for records until the first element changes, then write an output record ... but I would prolly be missing the last record.
Was trying to work with lookup tables, but I don't have unique keys.
Would it be feasible to use transforms purely for building up internal hashes and writing output from finalize?
Any other ideas, thoughts?
TIA,
Karl.
-
Hello,
Denormalize component does this.
Please sign in to leave a comment.
Comments 3