Calculating The Time Required To Update Statistics
Mar 6, 2008
We have a proc that runs nightly which updates stats. For each database it will run dbcc updateusage on each table, then then sp_updatestats on the database itself.
This process normally takes about an hour, but recently the time has gone up dramatically.
I am looking at ways of making the run faster. One suggestion was to split the work and have it run in parrallel, with multiple procs all doing the same job but on different databases. So for example, I could have a proc that would determine which databases would take the longest to process, and then call 3 different copies of our update proc, each getting a different set of databases to process.
If I went this route, how could I determine which databases would take the longest to complete this process? I'm investigating determining this from sysindexes' rowmodctr, or maybe just database size?
Any suggestions on how I could get an estimate of which dbs would be the most costly to update stats? This is for SQL Server 2000.
What is the unit of the numbers you get in the Time Statistics-part when running a query in Microsoft SQL Server Management Studio with Client Statistics turned on?
Currently I get mostly 0´s, but if I try and *** up a query on purpose I can get it up to around 30... Is it milliseconds or som made up number based on clockcycles or... ?
I would also like to know if it´s possible to change the precision.
AA Guyz i want to calculate class position of students from a table.Sample data is as follows...Roll # - Name - Marks1 - ABC - 602 - DEF - 603 - GHI - 574 - JKL - 555 -MNO - 506 -PQR - 53The query should return the following result.Roll # - Name - Marks - POSITION1 - ABC - 60 - 12 - DEF - 60 - 13 - GHI - 57 - 34 - JKL - 55 - 45 -MNO - 50 - 56 -PQR - 53 - 6I want query in MS SQL Server 2000
Hello, I am fairly new to this, but I need to write a query that calculates whether an order made a shipping cutoff based on creation time, and when it actually shipped, based on local customer time. All times in the table below are central time. Heres where it gets complicated for me. If the order was created after 8:00am local customer time the same day, and shipped prior to 5pm local customer time, it would be considered a pass, or a 1. If an order was created after 8am local customer time, BUT didnt ship until AFTER 5pm local customer time, it would be considered a failure or a 0. And to complicate things even more, if the order is created after 5pm local customer time, the order create time would be treated as it was created at 8am local customer time the following day local customer time, and would need to be shipped prior to 5pm local customer time that day. Would CASE be appropriate for this? Or am I thinking wrong? Any help would really be appreciated!!
I used SET STATISTICS TIME ON to get execution stats for a query. I found that the CPU Time was sometimes greater than the elapsed time. How is this possible? The query does not use any parallelism since I used the query option MAXDOP 1. Is the elapsed time wait time? Is the total execution time the sum of the CPU time and elapsed time?
I have a table called Tickets which contains ticket information for a machine. Each machine can have more than one ticket number opened at the same time. The ticket number contains start date/time and end date/time of the ticket. Thereefore the table looks something like this:
I want to be able to calculate total duration time(in hours) that EACH MACHINE had a ticket open...but here is the tricky part. The total duration time that a machine had ticket open has to encompas any tickets that may fall in the same time period. For example: If Machine A has a ticket open at 8:30 and the ticket is closed at 10:00. Meanwhile, Machine A had another separate ticket open at 9:30 which was closed at 10:30. In this case, the total duration time for this machine would be from 8:30 to 10:30 for a total of 2 hrs duration time.
Can anyone help me get started in tackling this problem or provide any examples?
Code: CREATE TABLE [dbo].[DigiTracker]( [DigitrackID] [int] IDENTITY(1,1) NOT NULL, [contact_id] [int] NOT NULL,
[Code]....
-- not accurate b/c it returns 0 for all TimePerPageSec SELECT DATEDIFF(second,min(createDt),max(createDt)) AS TimePerPageSec FROM DigiTracker GROUP BY contact_id, ZinePageNumber time spent (in seconds) between each record
Hi guys, I am having difficulty calculating the time duration between receiving process to shipping process. I have a table that consists of: Order#, Processes, Time_In, Time_Out. Order# can be 1, 2, 3, 4, 5. While at the same time Order# 1 can go through more than one process, i.e.: Receiving, VisualTest, MechanicalTest, ..., Shipping. Every Order# does not necessarily goes through all processes, but surely they will go through receiving process and shipping process. For each process we will have recorded time when the order# comes in and when it finishes with each process. I need to calculate the length of time from Time_In from Receiving to Time_Out in Shipping.
Hi,I have a table called Bookings which has two important columns;Booking_Start_Time and Booking_End_Time. These columns are both of typeDATETIME. Given any day how can I calculate how many hours are availablebetween the hours of 09.00 and 17.30 so a user can see at a glance how manyhours they have unbooked on a particular day (i.e. 8.5 hours less the timeof any bookings on that day), can this be done with a queryor do I have to work it out in my code?Thanks for your help
Hi guys, I am having difficulty calculating the time duration between receiving process to shipping process. I have a table that consists of: Order#, Processes, Time_In, Time_Out. Order# can be 1, 2, 3, 4, 5. While at the same time Order# 1 can go through more than one process, i.e.: Receiving, VisualTest, MechanicalTest, ..., Shipping. Every Order# does not necessarily goes through all processes, but surely they will go through receiving process and shipping process. For each process we will have recorded time when the order# comes in and when it finishes with each process. I need to calculate the length of time from Time_In from Receiving to Time_Out in Shipping.
I am contemplating creating a job to execute every 5 mins which will update index statisics if they are more than say 8% out. I would like to know what thoughts people have on this? i.e. pros and cons.
I like forward to what you have to say.
I have auto stats on. Our stats are often more than 10% out. At what level do you reckon the query plan might be effected by out of data stats?
It seems to me there are many ways to update statistics for a table. i.e. "sp_updatestats", "sp_recompile", "dbcc updateusage"
Can somebody tell me the difference between those commands and what's the best way for updating your statistics? Does reindexing update the statistics?
To update statistics for entire DB i have taken the script from under given link.But need to know the 1 : what is sample percent on update statistics 2 : will it be applicable for 2005 ?
script taken from : http://weblogs.sqlteam.com/tarad/archive/2006/08/14/11194.aspx
Hi All, I update statistics for three tables every day 2:00 AM and in the job we call one stored procedure and, in that stored procedure only three statements are writtern for update statistics Like: Exec('update statistics TBL1 with fullscan') Exec('update statistics TBL2 with fullscan') Exec('update statistics TBL3 with fullscan') And this job was working fine since many months but last two days its getting fail and it gives the error messages like : could not continue scan with NOLOCK due to data movement So could you help me what is the solution for this
I would like to know when we upgrade SQL Server 2000 database to SQL Server 2005 is it required to update the statistics even if we rebuild all the indexes or create new indexes?
I am planning to change our current UPDATE STATISTICS strategy, which is auto stats ON. Our database is terrabytes sizes and some tables with millions of rows with over 200 indexes in one table. Some of these indexes are not really used. Most of the tables are very small.
Droping and creating new indexes are quite often used in our environment. So static script may not help.
How can I identify most frequently used indexes in a table?
With the Microsft recommended auto stats ON, what are the best other practices I can include to improve the effeciency?
Any help would be apprecited. It would be realy great if any of you can share some scripts to generates dynamic scripts.
Thanks for setting up such a great site and forum.
Here is my problem:
I have a table like the following in SQL Server 2005:
order | taskid | main_person | temp_person | start_assign_date 1 | 3 | John | John | 2008-01-01 10:20:22 2 | 3 | John | Joe | 2008-02-05 15:20:22 3 | 3 | John | John | 2008-02-07 20:25:20 4 | 6 | Joe | Joe | 2008-01-01 10:20:22 5 | 6 | Joe | Mike | 2008-02-01 10:20:22 6 | 10 | Doug | Doug | 2008-01-01 10:20:22 7 | 7 | Russ | Russ | 2008-02-01 11:20:22 8 | 7 | Russ | Mike | 2008-02-08 12:20:22 9 | 7 | Russ | John | 2008-02-10 20:05:12
It was made to record who was in charge of a specific task at a specific time. Each task has its own main responsible person and some substitutes for that person as Temporary Persons (who did the task while main person was away). The Main Person's name is in the temp_person column when he is doing the task by himself.
I'd like to generate a report that shows: - in a specific time period - which persons were in charge of a specific task and - for how long
Something like this:
From 2008-##-## to 2008-##-## Task 3 - John - 15 days Task 3 - Joe - 5 days Task 6 - Joe - 18 days Task 6 - Mike - 2 days Task 10 - Doug - 20 days
I have some ideas to do that when there are both start and end dates for every record but I couldn't find a way to use the next assignment start date for each task, as the end date for its previous record (in that task group) to calculate the duration for that record.
I can group the tasks and users and put them in the chronological order but I can't indicate the next start date as the end date for the previous record (in specific task group) to use the date difference functions.
Hi all As you know when you run a piece of TSQL scrypt in Query Analizer at the bottom of the page , sql-server will show you the Time of completion of your code....is there any way to capture this time from SQL-Server environment and use it in the Front-End Application to inform the user?
I have a data set like so:UTC_TIME Timestamp NodeID Message FlagLineStation11/19/2005 10:45:07 1132397107.91 1 3 5 1028103411/3/2005 21:05:35 1131051935.20 2 3 5 1009104311/25/2005 21:12:16 1132953136.59 3 3 5 10371049I added the UTC_TIME column in as aconversion of the unix timestamp inthe TIMESTAMP column.Keeping things simple and straightforward, I need to be able tocalculate the difference from one record to the next (ordered byTIMESTAMP or UTC_TIME) and output the result into another column in thetable.NODEID is the unique id.First, what is the function to do so if, say, I only wanted tocalculate the difference between 2 records as just a basic SELECTstatement. That way I can answer quick question based on any one or twoNODEID's.Second, how would I further that to continually calculate (as statedabove)?WOuld this be a stored procedure? A trigger? A cursor?I am learning as I go here. Any help is greatly appreciated.R.
select TOP 10 rec_id,trans_id,number_id,card_no,message_id,trans_datetimefrom [dbo].[trans_log] order by trans_datetime desc101, 1,34343, 99999, 200, 2015-11-23 12:27:25.710101,2,34343,99999,210,2015-11-23 12:27:26.710102,3,43434,88888,200,2015-11-23 12:28:26.714102,4,43434,88888,233,2015-11-23 12:28:27.710expected result:34343,99999,datediff(ss,'2015-11-23 12:27:26.710','2015-11-23 12:27:25.710') --difference between row 2 and row 143434,88888,datediff(ss,'2015-11-23 12:28:26.714','2015-11-23 12:28:27.710') --
difference between row 4 and row 3difference between row 6 and row 5...In the above query, I always want to find the difference in transaction date time between second and first row in a moving window.I have the unique id as rec_id to compare the next row with the previous row.
On a SQL 7 sp 2 server, I have a database with about about 77,000 records, with automatic update statistics on inserting 1000 records took 43 minutes. With automatic update off, it took 23 minutes to insert the same 1000 records. On the same machine, I inserted 1000 records into 2 other databases with the same database structure and automatic update statistics on. On the second database, there are about 174000 records and it took 35 minutes to insert 1000 records. On the third database, there are about 93000 records and it took 19 minutes to insert 1000 records.
I have a query that retrieves a single record from searching on two tables. The statement goes like this... select sum(amount) from Table1 A union Table2 B on a.id = b.id where date < ### and date > #### and account = ###
As people are running a particular report, this statement is executed time and time again to pull up the numbers necesarry for the report. When the report gets slow, I can speed it up by updating the statistics. My concern is that I'm having to update the statistcs every hour; otherwise, the query becomes slow. I have noticed that users are inserting data while users are running the report on one of the tables listed above. I'm sure that's making it become more fragmented and ultimately slowing down the query. Do you have any suggestion on how I can make the union of these two tables faster? Or is there anything I could do to speed the query besides creating clusted indexes? Any help would be appreciated....thank you
I am maintaining a large table with millions of rows that has two non clustered indexes and data changing frequently, I need to keep the indexes fresh. Update Statistics runs much quicker than Reindex. What is the appropriate situation for each and why? Thanks in advance.
We are upgrading from sql 7 to 2000.During the upgrade process do we have to do a reindexing of all tables or will update statistics take care of that.
Or do we have to do both? What is the difference between reindexing and update statistics.
I have recently defragged my SQL server using INDEXDEFRAG. Can somebody please tell me how to update the statistics on all the tables? Thanks in advance.
Below is the script that I executed to defrag all the tables in my database if anyone needs this.
/*Perform a 'USE <database name>' to select the database in which to run the script.*/ -- Declare variables SET NOCOUNT ON DECLARE @tablename VARCHAR (128) DECLARE @execstr VARCHAR (255) DECLARE @objectid INT DECLARE @indexid INT DECLARE @frag DECIMAL DECLARE @maxfrag DECIMAL
-- Decide on the maximum fragmentation to allow SELECT @maxfrag = 20.0
-- Declare cursor DECLARE tables CURSOR FOR SELECT TABLE_NAME FROM INFORMATION_SCHEMA.TABLES WHERE TABLE_TYPE = 'BASE TABLE'
-- Loop through all the tables in the database FETCH NEXT FROM tables INTO @tablename
WHILE @@FETCH_STATUS = 0 BEGIN -- Do the showcontig of all indexes of the table INSERT INTO #fraglist EXEC ('DBCC SHOWCONTIG (''' + @tablename + ''') WITH FAST, TABLERESULTS, ALL_INDEXES, NO_INFOMSGS') FETCH NEXT FROM tables INTO @tablename END
-- Close and deallocate the cursor CLOSE tables DEALLOCATE tables
-- Declare cursor for list of indexes to be defragged DECLARE indexes CURSOR FOR SELECT ObjectName, ObjectId, IndexId, LogicalFrag FROM #fraglist WHERE LogicalFrag >= @maxfrag AND INDEXPROPERTY (ObjectId, IndexName, 'IndexDepth') > 0
-- Open the cursor OPEN indexes
-- loop through the indexes FETCH NEXT FROM indexes INTO @tablename, @objectid, @indexid, @frag
I am looking to run UPDATE STATISTICS for the first time, don't ask why it wasn't done prior please :(, on a set of large tables in our 346gb database whcih has been being populated with transactional data for the past 4 years. The tables contain 1.2, 35, 64, and 92 million rows. I have used the STAT_DATE function to determine that none of these tables have ever had update statistics run for them.
My question is how should I go about this process and what options should I be selecting when issuing the command? I assume that I must first run with the FULLSCAN paramater in order to initially generate statistics for the table then would assume that following this initial population I could run without any paramaters nightly against the tables in the database to keep statistics up to date. Any guideance you all could provide to a newb would be greatly appreciated.
