Unpivoting of table in defferent way due to performance

Hello,

I can’t help myself, so I’m turning here.

I have a table which has 97 columns and 12 000 000 rows.

I am working on a computer which has 32 GB RAM for Knime 4.7.7 I have dedicated 24 GB RAM.

The table contains historical changes i.e. each record is defined by one ID.

The workflow works for me but unfortunately it takes three days , for one above table . I have 100 such tables. Of course not all tables have so many columns and so many rows but still it is a larger number of tables.

Unpivot is good but my idea is to take each column in a cycle and evaluate it separately and at the end of the cycle combine it into one table .

The idea is to find out the most changing values in a given columns over a certain period.

Any idea?

Marek