Need some advice from experts on designing a flow. Create a service that will read a csv file which may contain anywhere over 6000+ rows of individual ids as shown in the sample below.

Need to read that file and go to oracle database and fetch a vname,vnumber,vid of each id in the csv and then go to document repository i.e. Oracle UCM and download all documents matching vname,vnumber,vid there can be >=0 documents for each vname,vnumber,vid and save them on a file system. UCM exposes a webservice to dowload the documents. Finally create a new csv appending the filenames that are downloaded for each id. Need to keep track of any errors but need to make sure to go over the whole ids in the csv to download the documents and skip in case of errors.

Need some advice on how to go about designing this as there may be over 6000+ rows in a csv file and looping it and hitting the database for each individual id and then hitting a UCM may be a bit expensive so open for any idea. How to go by designing this solution. Wondering if messaging can be helpful here or offloading process of getting the vname,vnumber,vid to pl/sql packages, creating staging tables etc.

Initial csv that contains ids:

**ID**
12345a
12s345
3456fr
we9795
we9797

Final csv output:

**ID        Files Downloaded from UCM**
12345a    a.pdf,b.doc,d.txt
12s345    a1.pdf,s2.pdf,f4.gif
3456fr    b.xls 
we9795
we9797    x.doc

Thanks