Reindex and Update Statistics History

  • Hi,

    Is there any way I could get the information like when the reindexing and update statistics was done on a particular database?

    Regards

    Sandhya

  • You can view when the stats were last updated for a particular set of stats by expanding the database in SSMS, expand the tables folder, then the statistics folder, and then right click on the set of stats and select properties.

    Gethyn Elliswww.gethynellis.com

  • Refer to:

    Index Related Dynamic Management Views and Functions (Transact-SQL)

    at:

    http://technet.microsoft.com/en-us/library/ms187974.aspx

    This code works in 2005, but I have not tested in 2008 ... give it a try if it does you have some of what you asked for. If it does / does not work, please post and then my knowledge will increase.

    SELECT 'Index Name' = i.name, OBJECT_NAME(i.id) AS 'Table Name',

    Case Indid

    WHEN 1 THEN 'Clustered'

    ELSE 'NonClustered'

    End 'Type',

    'Last Updated' = STATS_DATE(i.id, i.indid),rowmodctr AS

    '# Rows inserted deleted or updated', --, o.type

    i.keys

    FROM sysobjects o, sysindexes i

    WHERE o.id = i.id AND (o.type <> 'S' AND indid <> 0 AND indid <> 255)Gets date of last statistics update number of rows added, deleted or updated since last update

    Same caveat as above, but this will also report on Auto Created Statistic

    SELECT o.name AS Table_Name,i.name AS Index_Name,STATS_DATE(o.id,i.indid) AS Date_Updated

    FROM sysobjects o

    JOIN sysindexes i ON i.id = o.id

    WHERE xtype = 'U' AND i.name IS NOT NULL

    ORDER BY o.name ASC,i.name ASC

    Again tested on 2005 needs to be tested on 2008. Again asking you to report back if it does / does not work in 2008

    This will tell you the date and a lot more than what you would want to know about indexes.

    SELECT o.name as 'Table', i.name as 'Index', c.name as 'Column',

    'Index Type' =

    CASE

    WHEN PATINDEX('%_wa_sys_%',i.name) = 0

    THEN 'Index'

    ELSE 'Statisical'

    END,

    'Primary' =

    CASE

    WHEN (i.status & 0x800)= 0

    THEN 'No'

    ELSE 'Yes'

    END,

    'Clustered' =

    CASE WHEN (i.status & 0x10)= 0

    THEN 'No'

    ELSE 'Yes'

    END,

    'Unique' =

    CASE WHEN (i.status & 0x2) = 0

    THEN 'No'

    ELSE 'Yes'

    END,

    'Ignore Dup Key' =

    CASE WHEN (i.status & 0x1) = 0

    THEN 'No'

    ELSE 'Yes'

    END,

    'Ignore Dup Row' =

    CASE WHEN (i.status & 0x4)= 0

    THEN 'No'

    ELSE 'Yes'

    END,

    'No Recompute' =

    CASE WHEN (i.status & 0x1000000) = 0

    THEN 'No'

    ELSE 'Yes'

    END,

    'Computed' =

    CASE WHEN (c.iscomputed) = 0

    THEN 'No'

    ELSE 'Yes'

    END,

    'Nullable' =

    CASE WHEN (c.isnullable) = 0

    THEN 'No'

    ELSE 'Yes'

    END,

    i.OrigFillFactor AS 'Orig Fill Factor', i.rowcnt as 'Est.RowCount', i.reserved * cast(8 as bigint) as ReservedKB,

    i.used * cast(8 as bigint) as UsedKB, t.name as 'Column Type',

    'Precision' =

    Case c.xprec WHEN 0

    THEN ' '

    ELSE CAST(c.xprec as VARCHAR(3))

    END,

    'Scale' =

    Case c.xscale

    WHEN 0

    THEN ' '

    ELSE CAST(c.xscale as VARCHAR(3))

    END,

    c.Length as 'Length', 'Updated' = STATS_DATE(i.id,i.indid)

    from sysobjects o with(nolock) inner join sysindexes i with(nolock) on o.id = i.id

    inner join sysindexkeys k with(nolock) on i.id = k.id and I.indid = K.indid

    inner join syscolumns c with(nolock) on k.id = c.id and K.colid = c.colid

    inner join systypes t with(nolock) on c.xtype = t.xtype where o.xtype <> 'S' -- Ignore system objects

    Order By o.name, i.name

    If everything seems to be going well, you have obviously overlooked something.

    Ron

    Please help us, help you -before posting a question please read[/url]
    Before posting a performance problem please read[/url]

  • I just ran the script on my 2008 developer edition instance and it worked fine...as long as it answers the OP's question.

    Gethyn Elliswww.gethynellis.com

  • GRE (Gethyn Ellis)

    Thanks for the feed back. Edited my posting to add 2 additional scripts would you be so kind as to give them a test.

    Again I thank you ,and so will the original poster of the question and any others who stumble upon this forum.

    If everything seems to be going well, you have obviously overlooked something.

    Ron

    Please help us, help you -before posting a question please read[/url]
    Before posting a performance problem please read[/url]

  • bitbucket-25253 (3/16/2010)


    GRE (Gethyn Ellis)

    Thanks for the feed back. Edited my posting to add 2 additional scripts would you be so kind as to give them a test.

    Again I thank you ,and so will the original poster of the question and any others who stumble upon this forum.

    Both additional Scripts run OK on 2008 against the AdventureworksDW db on SQL2008

    The very first script (Script 1) returned 88 rows

    The 2nd script (Script 2) returned 78 rows -- for some reason I was expecting this to more than the first script?

    The final script (Script 3) returns 122 rows

    Gethyn Elliswww.gethynellis.com

  • GRE (Gethyn Ellis)

    Again many thanks .... now I have learned more and I hope the person asking the original question has also learned something new.

    You assistance in testing is greatly appreciated.

    If everything seems to be going well, you have obviously overlooked something.

    Ron

    Please help us, help you -before posting a question please read[/url]
    Before posting a performance problem please read[/url]

  • Sandhya-371593 (3/16/2010)


    Is there any way I could get the information like when the reindexing and update statistics was done on a particular database?

    Sandhya,

    Statistics are always created and updated with an index:

    SELECT [schema_name] = SCHEMA_NAME(T.[schema_id]),

    table_name = T.name,

    T.type_desc,

    index_or_statistics_name = S.name,

    is_auto_stats = S.auto_created,

    user_created = S.user_created,

    last_updated = STATS_DATE(T.[object_id], S.stats_id)

    FROM sys.tables T

    JOIN sys.stats S

    ON S.[object_id] = T.[object_id]

    ORDER BY

    T.[schema_id],

    T.name,

    S.stats_id;

  • It worked for me both on 2005 and 2008. Thanks so much !!!

  • Sandhya-371593 (3/17/2010)


    It worked for me both on 2005 and 2008. Thanks so much !!!

    No worries.

  • just for curosity...

    what about 2000? how to check the update?

    ----------
    Ashish

  • This code works in SQL 2000

    SELECT 'Index Name' = i.name, OBJECT_NAME(i.id) AS 'Table Name',

    Case Indid

    WHEN 1 THEN 'Clustered'

    ELSE 'NonClustered'

    End 'Type',

    'Last Updated' = STATS_DATE(i.id, i.indid),rowmodctr AS

    '# Rows inserted deleted or updated', --, o.type

    i.keys

    FROM sysobjects o, sysindexes i

    WHERE o.id = i.id AND (o.type <> 'S' AND indid <> 0 AND indid <> 255) --Gets date of last statistics update number of rows added, deleted or updated since last update

    This works in SQL 2000

    SELECT o.name AS Table_Name,i.name AS Index_Name,STATS_DATE(o.id,i.indid) AS Date_Updated

    FROM sysobjects o

    JOIN sysindexes i ON i.id = o.id

    WHERE xtype = 'U' AND i.name IS NOT NULL

    ORDER BY o.name ASC,i.name ASC

    And this works in SQL 2000 - (All tested using Query Analyser - Northwind DB

    SELECT o.name as 'Table', i.name as 'Index', c.name as 'Column',

    'Index Type' =

    CASE

    WHEN PATINDEX('%_wa_sys_%',i.name) = 0

    THEN 'Index'

    ELSE 'Statisical'

    END,

    'Primary' =

    CASE

    WHEN (i.status & 0x800)= 0

    THEN 'No'

    ELSE 'Yes'

    END,

    'Clustered' =

    CASE WHEN (i.status & 0x10)= 0

    THEN 'No'

    ELSE 'Yes'

    END,

    'Unique' =

    CASE WHEN (i.status & 0x2) = 0

    THEN 'No'

    ELSE 'Yes'

    END,

    'Ignore Dup Key' =

    CASE WHEN (i.status & 0x1) = 0

    THEN 'No'

    ELSE 'Yes'

    END,

    'Ignore Dup Row' =

    CASE WHEN (i.status & 0x4)= 0

    THEN 'No'

    ELSE 'Yes'

    END,

    'No Recompute' =

    CASE WHEN (i.status & 0x1000000) = 0

    THEN 'No'

    ELSE 'Yes'

    END,

    'Computed' =

    CASE WHEN (c.iscomputed) = 0

    THEN 'No'

    ELSE 'Yes'

    END,

    'Nullable' =

    CASE WHEN (c.isnullable) = 0

    THEN 'No'

    ELSE 'Yes'

    END,

    i.OrigFillFactor AS 'Orig Fill Factor', i.rowcnt as 'Est.RowCount', i.reserved * cast(8 as bigint) as ReservedKB,

    i.used * cast(8 as bigint) as UsedKB, t.name as 'Column Type',

    'Precision' =

    Case c.xprec WHEN 0

    THEN ' '

    ELSE CAST(c.xprec as VARCHAR(3))

    END,

    'Scale' =

    Case c.xscale

    WHEN 0

    THEN ' '

    ELSE CAST(c.xscale as VARCHAR(3))

    END,

    c.Length as 'Length', 'Updated' = STATS_DATE(i.id,i.indid)

    from sysobjects o with(nolock) inner join sysindexes i with(nolock) on o.id = i.id

    inner join sysindexkeys k with(nolock) on i.id = k.id and I.indid = K.indid

    inner join syscolumns c with(nolock) on k.id = c.id and K.colid = c.colid

    inner join systypes t with(nolock) on c.xtype = t.xtype where o.xtype <> 'S' -- Ignore system objects

    Order By o.name, i.name

    If everything seems to be going well, you have obviously overlooked something.

    Ron

    Please help us, help you -before posting a question please read[/url]
    Before posting a performance problem please read[/url]

  • Very nice Ron!

    Can't believe anyone is still using 2000 though 😉 😛

    (Please no-one take that statement seriously - thanks)

  • Thanks a lot. In my quick reference book now.

    unfortunately, some companies still using 2000 who still not able to decide either move to 2005 or 2008....lol :-D:-D:-D:-D:-D:-D

    ----------
    Ashish

Viewing 14 posts - 1 through 13 (of 13 total)

You must be logged in to reply to this topic. Login to reply