Showing posts with label analysis. Show all posts
Showing posts with label analysis. Show all posts

Monday, March 12, 2012

I think i killed Analysis Services (05), need a reset button

Hi,

I'm a cube rookie and have been deploying funky cubes while i'm cutting my teeth on AS. Unfortunately, AS has completely stopped responding, connecting to it in SQL Server Management studio or BI Studio just results in a hang. The service will stop and start without error in the Event log, but i can't interact with it any more. Restarting the PC doesn't help...

I tried to uninstall it, but setup is forcing me to uninstal SQL server with it, which isn't a good idea for me at this point. Is there any way to 'reset' AS?

Thanks!

Edited To Add:
Well, i've blown away any file i could find under this directory that appeared to have anything to do with my cube and i can now connect to AS...wish me luck!

C:\Program Files\Microsoft SQL Server\MSSQL.2\OLAP

Apparently in spur of the moment you killed too much :)

The files you needed to delete were under C:\Program Files\Microsoft SQL Server\MSSQL.2\OLAP\Data
Now you probably have no choice but uninstall and re-install the product again.

Run Remove for SQL Server, later in the uninstall wizard you should be able choose to remove Analysis Services only.

Edward.
--
This posting is provided "AS IS" with no warranties, and confers no rights.

|||Thanks Edward!

Fortunately I just left the detail out of my post, and only deleted the files under Data. So far things have been working well. I'm still learning a lot about how to put this stuff together, but the difference in performance between this and Access Pivot tables makes it all worth while.

Thanks!

Wednesday, March 7, 2012

I need some design help (time series analysis)

Hi,
Here is a fictious example that models my real life problem that I'm trying
to solve.
Lets say we have a bunch of baseball hitters and every day we give them an
hour to hit as many 'out of the park' home runs as they can hit. Now we want
to track home many they hit each day and watch whether their totals are
going up or down over time.
My fictious table design (as I see it) would be (with data)
PlayerId TotalHits Date
1 8 1/1/6
2 6 1/1/6
3 10 1/1/6
4 4 1/1/6
1 9 1/2/6
2 8 1/2/6
3 15 1/2/6
4 1 1/2/6
...
This may be poor design, please tell me if it is.
Anyway, from this design I could do a 'select playerid, sum(totalhits) from
xxx group by playerId' to see who has the most hits and look at each of the
top X hitters and see if they're in a incremental pattern or decreasing
pattern or just steady but high hits count type pattern.
Instead of that manual method I would like to figure a way to have the
system automatically tell me whose hit count is increasing over time and
whose are decreasing. and be able to break it down by week. month
quarter...
I'm guessing this might be analysis services in sql server, I've never used
them so I don't know. I run sql server 2000 btw.
If you have articles or references please point me in the right direction.
thanksHi,
the design looks ok, simple and straight forward, if you don=B4t want to
store more information here, you should be fine with that. Sure AS
could help you in some cases, but this should be just a simple query
here. Lets try to find a suggestion which will fit your needs. What do
you want to see at the end of the day, a list with playids only ? The
sums of their hits ? The difference between nother time period ? Which
time period do you want to investigate ? Try to give some sample
resultset that we can suggest something as an query for you.
HTH, Jens Suessmeyer
--
http://www.sqlserver2005.de
--|||"Jens" <Jens@.sqlserver2005.de> wrote in message
news:1151300420.928006.36970@.m73g2000cwd.googlegroups.com...
Hi,
the design looks ok, simple and straight forward, if you don´t want to
store more information here, you should be fine with that. Sure AS
could help you in some cases, but this should be just a simple query
here. Lets try to find a suggestion which will fit your needs. What do
you want to see at the end of the day, a list with playids only ? The
sums of their hits ? The difference between nother time period ? Which
time period do you want to investigate ? Try to give some sample
resultset that we can suggest something as an query for you.
HTH, Jens Suessmeyer
--
http://www.sqlserver2005.de
--
Ultimately I would like to see which players hit counts are on the rise and
which are on the decline. And yes I would like to be able to adjust the time
frames to analyze. So to start out I would like to see the query which would
show me the player id's whose hit counts are on the rise over the entire
time period stored in the database.
I can sort of see how the query could group by week (or month) but I don't
see how it can show me counts which are on the rise.
Thanks for your help.

Friday, February 24, 2012

I need some design help (time series analysis)

Hi,
the design looks ok, simple and straight forward, if you don=B4t want to
store more information here, you should be fine with that. Sure AS
could help you in some cases, but this should be just a simple query
here. Lets try to find a suggestion which will fit your needs. What do
you want to see at the end of the day, a list with playids only ? The
sums of their hits ? The difference between nother time period ? Which
time period do you want to investigate ? Try to give some sample
resultset that we can suggest something as an query for you.
HTH, Jens Suessmeyer
http://www.sqlserver2005.de
--Hi,
Here is a fictious example that models my real life problem that I'm trying
to solve.
Lets say we have a bunch of baseball hitters and every day we give them an
hour to hit as many 'out of the park' home runs as they can hit. Now we want
to track home many they hit each day and watch whether their totals are
going up or down over time.
My fictious table design (as I see it) would be (with data)
PlayerId TotalHits Date
1 8 1/1/6
2 6 1/1/6
3 10 1/1/6
4 4 1/1/6
1 9 1/2/6
2 8 1/2/6
3 15 1/2/6
4 1 1/2/6
...
This may be poor design, please tell me if it is.
Anyway, from this design I could do a 'select playerid, sum(totalhits) from
xxx group by playerId' to see who has the most hits and look at each of the
top X hitters and see if they're in a incremental pattern or decreasing
pattern or just steady but high hits count type pattern.
Instead of that manual method I would like to figure a way to have the
system automatically tell me whose hit count is increasing over time and
whose are decreasing. and be able to break it down by week. month
quarter...
I'm guessing this might be analysis services in sql server, I've never used
them so I don't know. I run sql server 2000 btw.
If you have articles or references please point me in the right direction.
thanks|||Hi,
the design looks ok, simple and straight forward, if you don=B4t want to
store more information here, you should be fine with that. Sure AS
could help you in some cases, but this should be just a simple query
here. Lets try to find a suggestion which will fit your needs. What do
you want to see at the end of the day, a list with playids only ? The
sums of their hits ? The difference between nother time period ? Which
time period do you want to investigate ? Try to give some sample
resultset that we can suggest something as an query for you.
HTH, Jens Suessmeyer
http://www.sqlserver2005.de
--|||"Jens" <Jens@.sqlserver2005.de> wrote in message
news:1151300420.928006.36970@.m73g2000cwd.googlegroups.com...
Hi,
the design looks ok, simple and straight forward, if you dont want to
store more information here, you should be fine with that. Sure AS
could help you in some cases, but this should be just a simple query
here. Lets try to find a suggestion which will fit your needs. What do
you want to see at the end of the day, a list with playids only ? The
sums of their hits ? The difference between nother time period ? Which
time period do you want to investigate ? Try to give some sample
resultset that we can suggest something as an query for you.
HTH, Jens Suessmeyer
http://www.sqlserver2005.de
--
Ultimately I would like to see which players hit counts are on the rise and
which are on the decline. And yes I would like to be able to adjust the time
frames to analyze. So to start out I would like to see the query which would
show me the player id's whose hit counts are on the rise over the entire
time period stored in the database.
I can sort of see how the query could group by week (or month) but I don't
see how it can show me counts which are on the rise.
Thanks for your help.|||"Jens" <Jens@.sqlserver2005.de> wrote in message
news:1151300420.928006.36970@.m73g2000cwd.googlegroups.com...
Hi,
the design looks ok, simple and straight forward, if you dont want to
store more information here, you should be fine with that. Sure AS
could help you in some cases, but this should be just a simple query
here. Lets try to find a suggestion which will fit your needs. What do
you want to see at the end of the day, a list with playids only ? The
sums of their hits ? The difference between nother time period ? Which
time period do you want to investigate ? Try to give some sample
resultset that we can suggest something as an query for you.
HTH, Jens Suessmeyer
http://www.sqlserver2005.de
--
Ultimately I would like to see which players hit counts are on the rise and
which are on the decline. And yes I would like to be able to adjust the time
frames to analyze. So to start out I would like to see the query which would
show me the player id's whose hit counts are on the rise over the entire
time period stored in the database.
I can sort of see how the query could group by week (or month) but I don't
see how it can show me counts which are on the rise.
Thanks for your help.

Sunday, February 19, 2012

I need help with Analysis Services

Hi.
I have a problem with the caculation of the dimensions in the Analysis
Services and I need help.
In a certain dimension I make some calculations with the "personalized member formulas". Some cells are modified after this calculations. After this, in another dimension, other calculations are made using again the "personalized member formulas". With this operations some cells that were modified in the previous dimension should be modified again, but they aren't. Is there a way to fix the sequence of the calculation of the dimensions? Do they get calculated at the same time or some get done before others? Once one of this cells are modified using those formulas is it possible to modify them again? In the Hyperion's Essbase software ther is an option to calculate some dimensions before others. Does Analysis Services has that option or something similar?.
I hope you can help me. Thank you very much.Are you using third party OLAP and if not what is the level of service pack on Analysis server?|||We are you using service pack 3 on Analysis server and the Microsoft tools for Analysis Services. (SQL Server 2000 SP3 + Windows 2000 Server + SQL Server Enterprise Manager; and Analysis Services SP3 + Analysis Manager)