Reputation
Top tag
Next privilege 200 Rep.
See reduced ads
Badges
1 7
Newest
 Scholar
Impact
~5k people reached

  • 0 posts edited
  • 0 helpful flags
  • 7 votes cast
May
11
comment Tune Performance by putting large tables into the KEEP Pool
kevinsky, this has been done. The storage is just so damn slow and we won't get a faster one. That means db file sequential read is so slow. We won't change an otherwise working application drastically and the original query has been optimized as much as possible.
May
11
accepted Tune Performance by putting large tables into the KEEP Pool
May
5
comment Tune Performance by putting large tables into the KEEP Pool
So if we do a SELECT /*+ FULL */ on the tables will they remain in their KEEP Pool forever if it is sufficiently large? Moreover, I don't think that RESULT_CACHE Hint will have huge effects, basically we cannot predict what the user might search for and all queries are hard parsed by purpose (because result set size varies so much). It will probably age out too soon. As our db file sequential reads are slow, I hope I can just keep the whole tables in memory all of the time.
May
5
asked Tune Performance by putting large tables into the KEEP Pool
Apr
8
comment How to find the most common phrases in a varchar field
It at least tokenizes for you and can give you token or 'hit' counts for phrases. Makes things much easier. What you are trying to do is in fact very computational intensive and quite academic. You can compare it to searching for substrings in human DNA. Google "String matching algorithms" and you will realize, that it won't be feasible to implement this in SQL imho.
Apr
8
comment How to find the most common phrases in a varchar field
Build a fulltext index and query it for the matches of every single token or connections of the tokens. Solving this in SQL will be far too complicated imho.
Apr
1
awarded  Scholar
Apr
1
comment Can Oracle have too much RAM?
Thank you, we will try huge pages on our Red Hat Enterprise Linux 5. I have not found any problem reports for that system on the web. So i assume, it will be kinda safe, but still leaves some sort of feeling.
Apr
1
accepted Can Oracle have too much RAM?
Apr
1
comment Can Oracle have too much RAM?
Thank you very much! I see your point now!
Apr
1
comment Can Oracle have too much RAM?
Hi, thank you. We are on an old Red Hat Enterprise Linux with 2.6.17 Kernel. Do we have to alter the Kernel? Can you please give me a link?
Apr
1
asked Can Oracle have too much RAM?
Mar
20
comment How to identify concurrency issues in Oracle 11g?
Indexes, Triggers? What happened during that update?
Mar
20
answered One to many to many
Mar
19
comment How does like query work with indexed table
You need to look at the index as a whole, for the whole table. It is some sort of file. If there's DML on the table the file slowly becomes fragmented.
Mar
19
comment How does like query work with indexed table
The index is for the whole table, so for all records in the table. DML will trigger changes to all indexes that are affected (e.g. multiple indexes containing the same columns).
Mar
19
comment How does like query work with indexed table
I don't really understand that questions. It's basically about disk space management. You want continguous segments for optimal performance and no duplicate or stale entries. However, this is not the case, you cannot compact/defragment larger indexes on every DML operation. It fragments over time and needs maintenance. And this is to my experience more common to fulltext indexes and has a larger impact on them, as they tend to grow larger.
Mar
19
comment How does like query work with indexed table
DML Operations can trigger updates on indexes (e.g. text changes, row deleted, new row inserted). These need to be written to the index and it's gotta be fast - you cannot wait on the index sometimes. Then the database decides to not cleanly insert the new value but rather put it somewhere else so it can be found. Or it deletes it but does not consolidate the space afterwards - and then you get index fragmentation.
Mar
19
comment How does like query work with indexed table
They are usually generated from the indexed texts (your columns in the FT index) and updated when there's DML on the table (DELETE/UPDATE/INSERT).
Mar
19
comment How does like query work with indexed table
Please look at my edit.