Hana Dictionary compression table access

[former_member92033]

Hi,

I would like to know how to access the dictionary tables created by HANA for compression. I would like to directly query this table.

Thanks,

Akshay

[lbreddemann]

You got that concept wrong. There is no dictionary table in the sense of an SQL table.

Maybe you want to read up a bit on how the column store works...

If you want to get all distinct values for a specific column, then the most efficient way to get consistent results is to simply do a

SELECT DISTINCT column_name FROM table_name

[former_member92033]

Hi Lars,

I was assuming that HANA dictionaries are stored in a similar format as HDI data. My application involves creation of a dictionary based on all columns of the entire database, which is why I want a way to access the already existing HANA dict. rather than creating one and burdening the system.

Essentially my application involves a step in which I need to represent the table..

Name   | City   | Country

Akshay | BOM | IN

In the below format

Akshay | Name

BOM     | City

IN         | Country

We have thought of solutions using a combination of triggers and procedures, but they are too complex. We also tried pushing the processing to the XSA layer but it introduces a lot of overhead. Also triggering the XSA script on row inserts is tricky.

If you have a solution for the same please do let me know.

Thanks,

Akshay T

[lbreddemann]

"on all columns of the entire database"

What does that even mean?

What about ID columns or date or numeric columns?

What about LONG value columns?

This sounds to me like a catch-all solution approach that actually requires a lot of detail work instead.

But no, there is no automatic mechanism that would go and scrape your whole DB to produce an arbitrary projection and transformation to provide column name and value as a record.

Not quite sure why you want to have a data model that has a lot of disadvantages but barely any advantages, but if your database has a decent amount of total rows and columns, you'll hit practical boundaries like the 2 bln row/partition rather soon.

[pfefferf]

Hello Akshay,

for that there exists no offical support and option for good reasons .

If you probably need that for technical reasons (bug analysis etc), I think you can ask your internal SAP colleagues.

Regards,

Florian