Re: tablespace calculating

the number datatype takes anything up to 21 bytes to store (for really really big numbers). The char(30) columns will take 30 bytes each. Therefore each row could contain up to

30*30 + 21 = 921 bytes of data. There is then a small overhead but to be honest given your example I'd assume each row took 1k to store. This overestimates and makes the maths easy your example gives a table size of 3m and 10k per day added. In practice in fact what I do, rather than follow the various convoluted calculation measures that there are is

1. estimate the current size of the data. If you already have it in some other form then so much the better just see how much room it currently takes.
2. Add 25% for growth.
3. allocate 50% of data size for index storage, any more and you've probably got your indexing wrong.

The reason being that any estimate will be wrong and after a while (say 3 months) you can see what the trend in storage actually is and adjust accordingly.

--
Niall Litchfield
Oracle DBA
Audit Commission UK


"Paul Meier" <stefan.porges_at_berlin.de> wrote in message
news:9opjuj$eb0q9$1_at_ID-109621.news.dfncis.de...


> Hi and  good morning and thanks to kevin for the first hint,

>

> could somebody please help. I dont know, how to calculate the needed

> tablespace.

> Sample table with 30 columns( 1 - column Number and 29 columns char(20) )

>                     initial data = 3000 rows

>                     average growth = 10 rows per day

>

> How many kb - tablespace should i furnish and why?

>

> Thanks a lot

>

> Paul

>

>