Hi,

If you have the data for more rows you could try to do batch inserts. 

http://stackoverflow.com/questions/12365992/sql-bulk-insert-statement

That reduces at least the roundtrips to the db...

Regards,
Sandor


mike matrigali <[email protected]> írta:
>I don't think there is anything specificly about a timestamp index vs
>another index that is bad.  All indexes will have an extra overhead
>at insert/delete/update time as they are implemented as a separate file
>maintaining an btree index on disk.  For your application it would 
>depend if the benefit to queries that use the timestamp collumn out 
>weigh the overhead of maintaining the index.
>
>Providing performance advice is hard.  About the only generic thing
>i can say about insert performance is to do as many inserts as possible
>in a single transaction so as to avoid I/O bound performance on the
>synchronous write of the commit record.  Also use and reuse prepared
>statements as possible.
>
>There are a lot of successful derby applications doing large numbers of
>inserts from evidence of past postings on derby user and derby developer 
>list.  To get more help I suggest you post more details to the list
>for those users to help.  Best would be to include a reproducible test
>case of the problem area, including source code.  That makes it the most
>likely you can get useful feedback.  Also including some basic 
>performance goals would be good - ie. hoping to get N inserts of rows 
>with the following DDL in M seconds made by P concurrent threads on
>X hardware with Y disk technology.
>
>On 1/8/2015 11:34 PM, kosurusekhar wrote:
>> Hi Folks,
>>
>> We have three tables where will have lot insertions per minute. In all three
>> tables we have TIMESTAMP columns, And we kept index for these timestamp
>> columns. This timestamp values always system time when the row is inserted.
>>
>> My doubt is whether maintaining the index tree for this kind of data
>> (timestamp) will be overhead for derby database. Because Timestamp contains
>> including seconds also right, I feel personally to create index structure &
>> maintaining this structure could be little overhead for derby server.
>>
>> Please correct me if I am wrong. In this scenario how to improve performance
>> while huge insertions happening into these tables.
>>
>> Thanks in Advance.
>>
>> Regards
>> Sekhar.
>>
>>
>>
>>
>> --
>> View this message in context: 
>> http://apache-database.10148.n7.nabble.com/Index-for-TIMESTAMP-COLUMN-tp143572.html
>> Sent from the Apache Derby Users mailing list archive at Nabble.com.
>>
>
>

Reply via email to