Top 10 SQL

Similar Questions

• TOP 10 SQL QUERIES

  Hello
  
  How can I find the 10 top sql queries which many high IO, CPU. in oracle db.
  I do it in a way that using the command TOP trying to get the PID then I get the sql query by applying the hash value in v$ sqlarea.
  
  Y there is it possible to get directly high consumed IO and CPU with see PID in TOP command.
  
  Thanks

  Hello

  try something along the lines of

  select c.* from
      (select disk_reads,
              buffer_gets,
              rows_processed,
              executions,
              first_load_time,
              sql_text
         from v$sqlarea
         where parsing_user_id !=0
         order by
            buffer_gets/decode(executions,null,1,0,1,executions) desc ) c
  where rownum < 11;
  
  select c.* from
      (select disk_reads,
              buffer_gets,
              rows_processed,
              executions,
              first_load_time,
              sql_text
         from v$sqlarea
         order by
            disk_reads/decode(rows_processed,null,1,0,1,rows_processed) desc ) c
  where rownum <11;
  

  or even

  --Top 10 by Buffer Gets:
  
  set linesize 100
  set pagesize 100
  SELECT * FROM
  (SELECT substr(sql_text,1,40) sql,
          buffer_gets, executions, buffer_gets/executions "Gets/Exec",
          hash_value,address
     FROM V$SQLAREA
    WHERE buffer_gets > 10000
   ORDER BY buffer_gets DESC)
  WHERE rownum <= 10
  ;
  
  --Top 10 by Physical Reads:
  
  set linesize 100
  set pagesize 100
  SELECT * FROM
  (SELECT substr(sql_text,1,40) sql,
          disk_reads, executions, disk_reads/executions "Reads/Exec",
          hash_value,address
     FROM V$SQLAREA
    WHERE disk_reads > 1000
   ORDER BY disk_reads DESC)
  WHERE rownum <= 10
  ;
  
  --Top 10 by Executions:
  
  set linesize 100
  set pagesize 100
  SELECT * FROM
  (SELECT substr(sql_text,1,40) sql,
          executions, rows_processed, rows_processed/executions "Rows/Exec",
          hash_value,address
     FROM V$SQLAREA
    WHERE executions > 100
   ORDER BY executions DESC)
  WHERE rownum <= 10
  ;
  
  --Top 10 by Parse Calls:
  
  set linesize 100
  set pagesize 100
  SELECT * FROM
  (SELECT substr(sql_text,1,40) sql,
          parse_calls, executions, hash_value,address
     FROM V$SQLAREA
    WHERE parse_calls > 1000
   ORDER BY parse_calls DESC)
  WHERE rownum <= 10
  ;
  
  --Top 10 by Sharable Memory:
  
  set linesize 100
  set pagesize 100
  SELECT * FROM
  (SELECT substr(sql_text,1,40) sql,
          sharable_mem, executions, hash_value,address
     FROM V$SQLAREA
    WHERE sharable_mem > 1048576
   ORDER BY sharable_mem DESC)
  WHERE rownum <= 10
  ;
  
  --Top 10 by Version Count:
  
  set linesize 100
  set pagesize 100
  SELECT * FROM
  (SELECT substr(sql_text,1,40) sql,
          version_count, executions, hash_value,address
     FROM V$SQLAREA
    WHERE version_count > 20
   ORDER BY version_count DESC)
  WHERE rownum <= 10
  ;
  

  You may need to play with the column formatting a bit to show the best results

  concerning

  Alan

  Published by: alanm on December 22, 2008 16:01

• top 10 SQL only for PeopleSoft

  Hello
  on HRCM91, tools 8.52 on Oracle DB 11 G R2,.
  How to query the Oracle DB to have 10 albums SQL only for PeopleSoft?
  
  There are actually in v$ sqlarea all text in sql. How to choose only those of PeopleSoft?
  
  Thank you.

  You can use CWA and ADDM reports. That would be advice you which ran for schema SYSADM (PeopleSoft) since the record names. you will find many other details as well.

  Thank you

• Top 10 SQL by physical reads

  Dear experts,
  
  Version of Oracle - 11.1.0.7
  
  Do we not have a way to draw 10 albums SQL (physical readings) for a period of 24 hours based on the AWR data or other points of view V$.
  
  Thank you

  DBA_HIST_SQLSTAT, all you can say is that was the user of the analysis.
  Not the user running.

  DBA_HIST_SQLSTAT is therefore insufficient for queries that are executed several times by several users, and there is not a lot of solution.

  You could have a glance in the underlying data from the ASHES - V$ ACTIVE_SESSION_HISTORY and DBA_HIST_ACTIVE_SESS_HISTORY which is respectively sampled data and a sample of a sample and it could give an indication of who is running.

  Published by: Dom Brooks on July 19, 2012 10:18
  Part of sentence was missing

• SQL Server 2012 and NUMA - practical implications

  Hello

  I read up on top of SQL Server and NUMA, with ESXi 5.1 in mind. I read that I should use the same NUMA architecture on hosts - this means in practice, I should use the same type, RAM speed, etc. when my guests? I want to know if my HP servers have the same NUMA architecture, but so far I was not able to find much information about it.

  I think just buying identical hardware to make sure that I understand.

  Thank you

  Hardware identical is certainly a big plus when your guest has related to NUMA optimizations, and your comments and request can take advantage of them.  With vNUMA, the optimization level and the benefits provided by the latter depends on the guest, as well as the material, so have the same host NUMA architecture will provide the most benefits.  The performance guide (http://www.vmware.com/pdf/Perf_Best_Practices_vSphere5.1.pdf)

  has additional details.

  -KjB

• SQL - CROSS Question

  Dear expert operator SQL Set,
  
  I have 5 queries that run in the database lists top 15 SQL ID, based on the highest - physical reads, buffer gets, executions Total, elapsed time and Cluster waiting respectively by each of the query. There will be a few common SQL identifiers that will appear in the result set of multiple queries. I would like to examine all 5 applications together and report SQL ID, sorted by number of appearances desc hight. How can this be achieved?
  
  If we consider 2 sets of results, a simple "INTERSECT" operator will do the job. What happened in my case? All thoughts... appreciate your help.
  
  Thank you.

  I just used the clause to generate data, given that I don't have something practical to use than real queries. If you want the sentence using the clause instead of views online, it's like:

  with reads as (
     select *
     from (select sql_id
           from dba_maint.DB_HIST_SQLSTATS
           where PARSING_SCHEMA_NAME in ('GLOGDBA','SNIOTM','OEM_READ') and
                 trunc(BEGIN_INTERVAL_TIME) between trunc(sysdate-1) and trunc(sysdate-1 )
           group by sql_id
          order by sum(disk_reads_delta) desc)
     where rownum < 16),
  gets as (
     select *
     from (select sql_id
           from dba_maint.DB_HIST_SQLSTATS
           where PARSING_SCHEMA_NAME in ('GLOGDBA','SNIOTM','OEM_READ') and
                 trunc(BEGIN_INTERVAL_TIME) between trunc(sysdate-1) and trunc(sysdate-1 )
           group by sql_id
           order by sum(buffer_gets_delta) desc)
     where rownum < 16)
  select sql_id , count(*)
  from (select sql_id from reads union all
        select sql_id from gets)
  group by sql_id
  order by 2 desc;
  

  Each as block is actually a view online, to separate from the main logic of the request for clarity (personal or lifestyle preference :-)))

  You could also easily do without the with as:

  select sql_id , count(*)
  from (select *
        from (select sql_id
              from dba_maint.DB_HIST_SQLSTATS
              where PARSING_SCHEMA_NAME in ('GLOGDBA','SNIOTM','OEM_READ') and
                    trunc(BEGIN_INTERVAL_TIME) between trunc(sysdate-1) and trunc(sysdate-1 )
              group by sql_id
              order by sum(disk_reads_delta) desc)
        where rownum < 16
        union all
        select *
        from (select sql_id
              from dba_maint.DB_HIST_SQLSTATS
              where PARSING_SCHEMA_NAME in ('GLOGDBA','SNIOTM','OEM_READ') and
                    trunc(BEGIN_INTERVAL_TIME) between trunc(sysdate-1) and trunc(sysdate-1 )
              group by sql_id
              order by sum(buffer_gets_delta) desc)
        where rownum < 16)
  group by sql_id
  order by 2 desc;
  

  John

• Difference between SQL and PL/SQL

  Hello
  
  Wrong if I say:
  SQL only access the database (under the hood)
  PL/SQL is a layer on top of SQL.
  
  Correct me if I'm wrong
  
  Thank you

  user12045475 wrote:

  Try to understand, it is true what I write below?

  What exactly is BC ? Cache buffers?

  SQL:
  Select *...
  data-online BC-online client memory file

  All SQL are analyzed as a SQL cursors. A handful of customer for this SQL cursor is returned to the caller. In the APIs such as ODBC, it is a statement handle. In PL/SQL, this is called a variable from PL/SQL cursor for example.

  The client now called the interface extraction (using this handle) for this SQL cursor. The data moves from the (buffer cache/LMS) database to the client variables that will receive this output cursor.

  If the client is remote, the data moves over TCP. If the client is local, that data travels via IPC (since the process of the Oracle server to the client process).

  If this customer is the PL/SQL, the data moves from area (SGA) memory to another (PGA) within the same server process. This is the fundamental reason why PL/SQL will always be better than other clients as the travel itinerary data from the SQL cursor for the client variables/extraction buffer is very short.

  Insert * select
  data-online BC-online data file file

  Same process as above. Only, the customer data from the database. It retrieves data from a table and inserts the data into another table - in the same cursor. The SELECTION of this slider component sends not all selected customer lines.

  This highlights a fundamental approach to Oracle called SQL maximize and minimize the PL/SQLdevelopment.

  The SELECT could be done via a dedicated slider. With the client by pulling all lines so that the database in the memory of the client. So, the INSERT could be done (also a dedicated slider) and the same data that is pushed back to the database.

  So for this approach, data will be make to the customer, then from the client to the database database. It's quite a detour for travel. The same same thing could be done using a simple SQL cursor (as opposed to two dedicated sliders). And this slider could make reading and writing lines, without sending the data via a detour through the memory of the client.

  The underlying concept is to optimize the use of SQL - solve both the problem of data processing in SQL itself. And if it is not really feasible using only SQL, then consider using the PL/SQL.

  PL/SQL:
  Select * in...
  data file => BC-online PGA? -online client memory

  See the answer in the upper part of the 1st sample you posted. This is no different. A SQL cursor is created (called a PL/SQL implicit cursor). The output of this slider is extracted and stored in local client variables (which reside in the PGA for PL/SQL code).

• AWR time model statistics vs SQL Statistics

  Looking at an AWR report took on a unique snapshot interval, I note the following:
  
  In time model statistics, DB time is listed as 161,47 seconds, with sql execute represents the time of 127,83 seconds.
  
  When I DBA_HIST_SQLSTAT to queries on the same cliché and summarize all the elapsed_time_delta of each SQL stored in AWR, I find that it's 132,05 seconds, even after filtering of all PL/SQL (command_type 47) and each SQL with parsing_schema_name SYS.
  
  If anything, I expect this sum less than that contained in the statistics of the model time, thinking that only the top n sql is included in DBA_HISSQLSTAT.
  
  Does anyone know where my reasoning was an error?

  KDeWeerd wrote:
  Detailed analysis of an environment, trying to give meaning to all the numbers and to explain things to people doesn't make not easier if you can not all numbers that collects Oracle corresponding to some how.

  Yes, I understand. Here you can not because the statistical statement have no knowledge of what what call. This is probably why Oracle introduced the chronological model: it gives information that you can't derive from other sources.

  Any timed operation will be buffer at most 5 seconds of time data. This means that if a timed (such as the SQL execution) takes a long period of time to perform, the data published to this point of view lack maximum 5 seconds of the time accumulated for the operation.

  Can't really make sense of interpretation which

  means that the value may be missed for 5 seconds for each running of instruction.

• in the dynamic chain of the clause

  Select all_group_id in the iolm_subelementgrp where type = "connect";

  ALL_GROUP_ID

  -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------

  3,4,5

  requirement:

  Select * from iolm_linkthresholdgrp

  where group_id in (3,4,5);  -How can I put values 'to' article like this dynamically, which is a discount on top of SQL?

  Select * from iolm_linkthresholdgrp

  where group_id in (select all_group_id in the iolm_subelementgrp where type = 'Connect');

  ORA-01722: invalid number

  01722 00000 - "invalid number."

  * Cause:

  * Action:

  Happy!

  Select regexp_substr('1,20,300','[^,]+',1,level)

  of the double

  connect by level<=>

  First see how it works!

  It should work-

  Select *.

  of iolm_linkthresholdgrp

  where group_id in)

  Select regexp_substr(all_group_id,'[^,]+',1,level)

  of iolm_subelementgrp

  When type = "sign in".

  connect by level<=>

  );

  I hope this helps!

• SQL_ID in different tables

  Hello
  on the 11g R2,.
  I wonder if there is no common SQL_ID in the following tables:
  

  select sql_id from DBA_HIST_SQLTEXT
   intersect 
  select sql_id from DBA_HIST_SQLSTAT 
  intersect 
  select sql_id from DBA_HIST_ACTIVE_SESS_HISTORY;
  
  no row selected

  If yes there is no SQL_ID located at least two of these tables what would be the reason?
  
  Thank you

  If yes there is no SQL_ID located at least two of these tables what would be the reason?

  You combine two tables AWR with a table of repository of ASHES.

  It is reaonable to require that all instructions in DBA_HIST_SQLSTAT are stored in DBA_HIST_SQLTEXT.
  Here are the tables in AWR and it is reasonable that any sql statement that is in the top N should have the preserved text.

  But DBA_HIST_ACTIVE_SESS_HISTORY is the repository of ASHES.
  The buffer in memory V$ ACTIVE_SESSION_HISTORY is once every second 'snapshot' of all active sessions.
  DBA_HIST_ACTIVE_SESS_HISTORY is once every 10 second sample of this snapshot.

  If SQL N albums directions are a snapshot of the active sessions every 10 seconds so is down to coincidence and circumstance and not something you should read too much into.

  Bugs despise or other oddities, it just suggests that your top N SQL instructions do not have tend to be long.

• New to Oracle DBA

  Recently I have been thrust into a role of DBA (kicking and screaming). My supervisor put me through the Oracle Sys Admin I & II workshops, but it's a starting point; It seems that many of the features that have been addressed are not used in our environment (ASM, AMM, RMAN, EM) that makes my life less funny!
  My job now is to monitor the database for all performance issues, the prevention of proactive failure, the backup and restore (BU / R is managed by NetBackup/BCV with Symmetrix disk mirrored between).
  
  I was an admin systems for many years, and I've been able to generate scripts a few sanitary controls outside the DB; at the level of the BONE. I'm starting to look at some of the views v$ that allow me to see what is happening in our system.
  I have a few generalized questions, I would ask (if I need a separate thread for each of the questions let me know and I'll repost):
  1. How can I determine what features are used by the database (outside the features mentioned above)?
  2. what should be followed on a base < interval > (hourly, daily, monthly)?
  3. what should I look for when I followed him? (right now I look at the alert log every day for entries "ORA" - and send alerts...)
  
  I'm sure I'll have many more questions, but these are ones I have on my mind right now.
  I apologize for the generality of my questions, I'm sure we'll get more accurate that the discussion continues.
  
  We had the following:
  Oracle 11 g 2 HA (Veritas Cluster NOT oracle cluster or grid control) mode running on Solaris 10...
  Thanks in advance for helping me understand this complex beast!

  Hello

  1. How can I determine what features are used by the database (outside the features mentioned above)?
  A >

  Under query will give the features that are currently enabled on your database.
  Select * option $ v where the VALUE = 'TRUE '.

  This will give you the details usage report features db if used

  Select * from DBA_FEATURE_USAGE_STATISTICS where CURRENTLY_USED = 'TRUE '.

  2. what should be followed on a basis (hourly, daily, monthly)?
  A >
  S/n should be followed always schedule/carefully, but you can't be human control grid or cronjobs (with automatic e-mail alerts) should be programmed.
  AWR/ADDM features are available that allows you to go back in time & see what was going on in DB giving excellent statistics of the DB

  3. what should I look for when I followed him? (right now I look at the alert log every day for entries "ORA" - and send alerts...)
  A >
  As grid needs licenses the simplest to be a SYSTEM administrator is to make some shell scripts and place a few jobs long life in cron
  This cron so captured ORA-, err, corrupted fail, warning, down, stop channels alerts log
  Apart from this you can also have scripts in place for session locking/blocking in DB, cpu/memory/IO TOP consuming SQL at the request to ship you once use Server metrics goes high
  One of the most important is alert, space free OS space, tablespace space free sql must always be runing in cron every min

  Thank you
  Ajay more
  http://moreajays.blogspot.com

• Run to analyze very small... already implemented standard tips

  Hi all
  I have on my Oracle DB performance problems, checking the AWR I noticed execution parse is very low. I looked into the present and other forums to find a possible root cause for her, but it seems still miss me something. I have already implemented the standard stuff, like optimizing SQL with bind, checking the size of the shared memory and using 'FORCED' as a value of cursor variables. However, I did get no improvement.
  
  AWR shows:
  
  Instance efficiency percentage (target 100%)

  Buffer Nowait %:          100.00               Redo NoWait %:     100.00
  Buffer Hit %:               99.96               In-mem Sort %:     100.00
  Library Hit %:               96.36               Soft Parse %:     98.74
  Exe. to Parse %:          9.38               Latch Hit %:     96.86
  Parse CPU to Parse Elapsd %:35.17               Non-Parse CPU:     98.41

  My results of the Shared Pool are:

  Memory Usage %:                55.11     55.71
  % SQL with executions>1:     77.53     34.35
  % Memory for SQL w/exec>1:     60.21     17.57

  Which means (I guess) maybe I lose the memory, but not that I'm not short on this subject.
  
  When I look in the top ten SQL sorted by Parse calls, I find statements like:

  1) select to_char(sysdate, 'YYYY-MM-DD hh24:mi:ss') into :b0 from DUAL 
  2) select lib_infospec into :b0 from table where ((field1=:b1 and field2=:b2) and field3='X')
  3) select SOME_THING.item into :b0 from DUAL
  4) select something from table where ((field1=:b0 and field2=:b1) and field3=:b2) 

  For each of them, "Analyze calls" is very close to "Executions".
  
  My questions:
  
  -I'm looking for the correct indicators? Do you have any other suggestions?
  -Why the two select statements double is always re-analysis? which can be avoided? How?
  -Since my cursor_sharing = force, why field3 = 'X' instead of a host variable in SQL statement nr.2?
  -Why SQL instruction nr. 4, which seems optimized for me, is re-arsed every time?
  
  Unfortunately, I can't post the full CWA, but I'll be happy to provide any additional information.
  
  Thanks in advance for your help,

  Welcome to the forum.

  Do you or developers use prepared statements?

  http://asktom.Oracle.com/pls/asktom/f?p=100:11:0:P11_QUESTION_ID:952783600346103200 #1643762500346954286

• Order By with null values

  Is it possible under the ORDER BY clause to not have null values climb on top
  
  Examples:
  
  SQL code...
  ORDER BY
  BL / / DESC
  
  ~*~*~*~*~*~*~*~
  RESULT
  bl*
  (null)
  8675309
  --------------------------------------
  ORDER BY
  BL CSA
  
  RESULT
  bl*
  (null)
  8675309
  
  
  Any thoughts?

  Yes, you can

  select * from emp  order by comm NULLS LAST  ;
  
  OR 
  
  select * from emp  order by comm NULLS FIRST;
  

  SS

• Obtain specific registration

  In SQL Server, I can do something like the following:
  
  Select top 1 * from [table] by id asc;
  
  He would return the first record in the table after the results had been ascending order. I know I can use ROWNUM to get something to use command 'top' of SQL Server, but I remember hearing that Oracle will ignore my ORDER BY clause and simply return the first record, regardless of the order. What is the best way to get the same result that SQL Server would return?
  
  In addition, where it's important, it will be to a 9i database.

  Use online viewing:

  select *
  from (select * from table order by id asc)
  where rownum =1;
  

• SQL server installation is either corrupt or has been tampered with. Please uninstall and run again together at the top.

  How can I fix this error? Thank you.
  SQL server installation is either corrupt or has been tampered with. Please uninstall and run again together at the top.

  Hi DMarie125,

  The question you have posted is better suited to the MSDN forums. You can ask your question in the help link.

  http://social.msdn.Microsoft.com/forums/en/category/SQLServer/

  Hope the helps of information.