155 reputation
7
bio website
location Kolkata, India
age 24
visits member for 2 years, 2 months
seen May 28 at 7:59

I am passionate about programming. I am entirely self-taught, having done my bachelors in Electrical Engineering, so I often have problems regarding my projects. What better site to answer my queries than stackoverflow?


Sep
24
awarded  Autobiographer
Apr
1
awarded  Popular Question
Sep
18
awarded  Commentator
Sep
18
comment How are multiple indexes used in a query by MySQL?
Thanks. One question, cardinality of column is important in deciding whether to apply index on a column. Can you please explain why? And why would cardinality play a role in deciding the ordering of the columns here? If you can update your answer with a detailed explanation, I would be really grateful.
Sep
18
accepted How are multiple indexes used in a query by MySQL?
Sep
18
asked How are multiple indexes used in a query by MySQL?
Sep
14
accepted Aren't two writes required to update a clustered index record
Sep
13
comment Aren't two writes required to update a clustered index record
Ummm, what is meant by that?
Sep
13
comment Aren't two writes required to update a clustered index record
I didn't quite understand your last statement. If you look at the ppt at google.co.in/… , you will find that when 5 is inserted, the 8 in the leaf level is pushed to the non-leaf level, effectively creating two copies of it.
Sep
13
comment Aren't two writes required to update a clustered index record
No, even there it says the that there are two different writes, not modification of object in two places.
Sep
13
comment Aren't two writes required to update a clustered index record
@AlexKuznetsov Have a look at the article at technet.microsoft.com/library/Cc917672 under the section Test 1: INSERT Performance.
Sep
13
awarded  Editor
Sep
13
revised Aren't two writes required to update a clustered index record
added 815 characters in body
Sep
13
comment Aren't two writes required to update a clustered index record
What do you mean by 'usually'? The new key has to be put into the index structure also, either inside an existing node without overflowing (lucky enough not to cause a cascade), or overflowing, in which further splits are needed, which might cascade up.
Sep
13
comment Aren't two writes required to update a clustered index record
Yeah, for indexes too. The row will get inserted into the correct position on the leaf level, but the new key value has to be put into the tree structure, right? For example look at the examples of inserting keys into a B+ tree at mec.ac.in/resources/notes/notes/ds/bplus.htm, each time the tree gets restructured.
Sep
13
asked Aren't two writes required to update a clustered index record
Sep
1
awarded  Student
Sep
1
accepted Need for reaching data through clustered index with a non-clustered index
Sep
1
awarded  Scholar
Sep
1
comment Need for reaching data through clustered index with a non-clustered index
Oh, you meant that all NIC's have to be updated. Sure. I though that you meant all index rows within the NIC have to be updated. That's why I said only the relevant index rows have to be updated. But these will of course add up cumulatively. Another question, the article you pointed to, says that if the NCIs hold row-ids, that's space wastage, for it's useless data. But when the NCIs are pointing to to the node, they all are containing a pointer to the node-id anyway, so where is the economy of space?