I have a table in MS SQL Server 7.0 Like this:

Id Cat_Name Definition Image
101 A1 --------------- *************
102 B1 ----------- *************
101 A2 ------------------ *************
103 C1 ----------------- *************
101 A3 --------------- *************
101 A4 ------------------ *************
102 B2 -------------------- *************
101 A5 -------------- *************
101 A6 ----------------- *************


From above table, How can I select 3 – 5 records where iId is 101. Means result should be:
<<101>><<A3>>----------------<<***********>>
<<101>><<A4>>----------------<<***********>>
<<101>><<A5>>----------------<<***********>>

might be able to help you with this but need a few questions answered.
1) I am assuming that ID+Cat_Name is unique, in other words, there cannot be two 101 A1 entries. is this correct?
2) what are specifications for the Cat_Name field? Is it always exactly two characters(one letter and then one number) or can it be one letter followed by any integer, etc.
3) you said 3-5 rows returned. Do you want that number random as well, or can it be the same number of rows returned everytime as long as that number is between 3-5?