Thats true Lars, the ECC compression seems far too much better, not so encourgaing from HANA perspective. Our compression on HANA side has been set up optimally. I did go through these presentations a while ago, the hidden columns are the biggest culprits actually across all our tables.
For the same table ANLP(50 GB in HANA):
$trexexternalkey$ occupies 27 GB and
Concat Attributes:
$ANLN1$ANLN2$BUKRS$ - 6 GB
$ANLN1$ANLN2$MANDT$ - 2.5 GB
$ANLN1$ANLN2$ - 2.5 GB
$ANLN2$BUKRS$ - 1.2 GB
Thats 40 GB out of the 50 by internal columns.
I know we can get rid of the concat attributes but they will get created again when the users run the reports again isn't it ? Do you know if there is a better way of compressing these HANA tables, not always an easy pill to digest when I look at the compressed size in ECC and compare it to HANA !!