Is the command Analyze table tbl compute statistics a DDL or DML? Intuitively, it seems to be neither.
When i have this command in a .sql file do i need to do :
execute immediate 'Analyze table tbl compute statistics'
I have a smiliar question about the command: GRANT DELETE, INSERT, SELECT, UPDATE ON tbl to user
UPDATE Oracle says that both grant and analyze are Data Definition Language (DDL) statements. They apparently do not make a distinction between DDL and Data Control Language (DCL).
If executing from within PL/SQL, then either execute immediate or DBMS_SQL would be needed.
Also, "Do not use the COMPUTE and ESTIMATE clauses of ANALYZE to collect optimizer statistics. " (10gR2) "For the collection of most statistics, use the DBMS_STATS
package. ... Use the ANALYZE statement (rather than DBMS_STATS)
for statistics collection not related to the cost-based
optimizer." (11g R2) Analyze table is deprecated for gathering optimizer statistics, though still usefull for other things. Use DBMS_STATS instead. (I linked to the online Oracle documentation for 10g R2. However I've been having trouble with Oracle's documenation site the last few days, with the 10g R2 documents disappearing and then reappearing.)
Related
I'm working on improving Oracle tables' performances.
The table that I have been working on created with 'monitoring' and 'logging' clause. These two, decreasing performance of query and I need to change monitoring to nomonitoring(without dropping tables).
This is working well:
alter table some_table nologging
But to alter monitoring to nomonitoring I use;
alter table some_table nomonitoring
query executes without any errors but there is no change in table structure.
I've been researching on internet for days and also as I saw here there is no such topic for my specific problem.
Thanks in advance.
The monitoring/nomonitoring options are deprecated and are no longer used in Oracle.
Quote from the Oracle 11.2 manual
Formerly, you enabled DBMS_STATS to automatically gather statistics for a table by specifying the MONITORING keyword in the CREATE (or ALTER) TABLE statement. Starting with Oracle Database 11g, the MONITORING and NOMONITORING keywords have been deprecated and statistics are collected automatically. If you do specify these keywords, they are ignored.
(Emphasis mine)
I have some question about Postgres, I have used dbms_stats.gather_table_stats for performance optimization in Oracle. I would like to switch our database from Oracle to Postgres, therefore, I want to achieve same feature on Postgres also. I searched internet whether there is some equivalent feature existing in Postgres with dbms_stats.gather_table_stats in Oracle. The only I found was EXPLAIN, VACUUM something like that. I think these are already existing in Oracle with same name. but I can't find proper ones for dbms_stats.gather_table_stats. I am spedning a lot time on it, if you guys have some advice, could I get some?
The GATHER_TABLE_STATS procedure of DBMS_STATS package collects statistics of the specified table in Oracle.
In Postgres, we use ANALYZE for the same purpose.
ANALYZE collects statistics about the contents of tables in the database, and stores the results in the pg_statistic system catalog. Subsequently, the query planner uses these statistics to help determine the most efficient execution plans for queries.
Can anyone suggest how to convert CONNECT BY Oracle query into Greenplum. Greenplum doesn't support recursive queries. So, we can not use WITH RECURSIVE. Is there any alternate solution to re-write the below query.
SELECT child_id, Parnet_id, LEVEL , SYS_CONNECT_BY_PATH (child_id,'/') as HIERARCHY
FROM pathnode
START WITH Parnet_id = child_id
CONNECT BY NOCYCLE PRIOR child_id = Parnet_id;
There are ways to do this but it will be a one-off per query. You will need to create a function that loops through your pathnode table and "return next" to return each row. You can search on this site to find examples of doing this with PostgreSQL 8.2.
Work is happening to rebase Greenplum to PostgreSQL 8.3, 8.4, and so on. Those later PostgreSQL versions support "with recursive" which is the ANSI SQL way to write your SQL but Greenplum doesn't support it yet. When it does get supported by Greenplum, I don't think it will perform all that well. The query will force looping and individual row lookups. This works great in an OLTP database but not so well for an MPP database.
I suggest you transform your data in Oracle with a VIEW and then just dump the view to a file to load into Greenplum. The DDL of having a self-referencing, N-level table will never be a good idea in an MPP database.
In Oracle I came across two types of insert statement
1) Insert All: Multiple entries can be inserted using a single sql statement
2) Insert : One entry will be updated per insert.
Now I want to insert around 100,000 records at a time. (Table have 10 fields with includes a primary key). I am not concerned about any return value.
I am using oracle 11g.
Can you please help me with respect to performance which is better "Insert" or "Insert All".
I know this is kind of a Necro but it's pretty high on the google search results so I think this is a point that worth making.
Insert All can give dramatic performance benefits if you are building a web application because it is a single SQL statement that requires only one round trip to your database. In most cases although far from all cases. the majority of the cost of a query is actually latency. Depending on what framework you are using, this syntax can help you avoid unnecessary round trips.
This might seem incredibly obvious but I have seen many, many production web applications in large companies that have forgotten this simple fact.
Insert statement and insert all statement are practically the same conventional insert statement. insert all, which has been introduced in 9i version simply allows you to do insertion into multiple tables using one statement. Another type of insert that you could use to speed up the process is direct-path insert - you use /*+ append*/ or /*+ append_values*/(Oracle 11g) hints
insert /*+ append*/ into some_table(<<columns>>)
select <<columns or literals>>
from <<somwhere>>
or (Oracle 11g)
insert /*+ append_values*/ into some_table(<<columns>>)
values(<<values>>)
to tell Oracle that you want to perform direct-path insert. But, 100K rows it's not that many rows and conventional insert statement will do just fine. You wont get significant performance advantage using direct-path insert with that amount of data. Moreover direct-path insert wont reuse free space, it adds new data after HWM(high water mark), hence require more space. You wont be able to use select statement or other DML statement, if you has not issued commit.
To use FORALL you would need PLSQL tables.
This process is quite fast.
You can also choose the table to have NO LOG option which would speed the process up during inserts.
Does oracle (11g at least, but it would be nice if 10g too) have some statistics on DML been performed?
What I expect to get is: for table A in schema B it was performed I inserts, J updates, K deletes from date1 to date2
You can use the monitoring feature and the dba_tab_modifications table to get that information. See for example here or here.