Since the data is split, I did this to have a single file to work on.

cat testData_tabular_chunk5.csv testData_tabular_chunk6.csv testData_tabular_chunk7.csv > testDataFull.csv

Since my CV and LB scores are not correlated at all, I just wanted to make sure.

PS1: Data this small does not seems to need to be split
PS2: Submisison could include the reference ID, that would remove any uncertainty

Thanks.