Where oracle records type LOB data?
I am new to oracle, so I don't know how the LOB data, are recorded as in postgresql all LOB is saved in a table (pg_largeobject) and all tables with LOB columns have an id for this "main table".
If oracle except LOB in a table and all the tables there just an id to the 'main' table or each table has its own copy of the lob.
Welcome to Oracle!
Oracle database stores the LOBs in the form of segments in a tablespace.
Suppose you have a table with a column lobcolumn lobtable, segment of the table is stored in the tablespace of the SCT and also his JOB. You can set a different tablespace to store the LOB segment if you wish.
Please take a look at the documentation.
http://docs.Oracle.com/CD/B28359_01/AppDev.111/b28393/adlob_tables.htm#i1006378
Kind regards
Franky
Tags: Database
Similar Questions
-
I'm not so familiar with LOB and was hoping someone could shed some light for me.
I am under Oracle 11.2.0.2 EE and made an interesting discovery of this new database I load.
First of all, I discovered that I have a table which is about 7.4 G, but there two columns LOB that when I have a dba_lobs query, I found that they contains 365 G of lobs and the table itself has 22 G of LOB - don't know what the difference is.
Now, let's look at the table these LOBS belong...SQL> 1 select segment_name, round(sum(bytes)/1024/1024/1024,1) as "SIZE" , segment_type 2 from dba_segments where owner = 'ARADMIN' 3 group by segment_name, segment_type 4 having round(sum(bytes)/1024/1024/1024,1) > 1 5* order by 2 / SEGMENT_NAME SIZE SEGMENT_TYPE -------------------------------- --------------- ------------------ . . . SYS_LOB0000077517C00027$$ 4.2 LOBSEGMENT SYS_LOB0000210343C00029$$ 4.4 LOBSEGMENT SYS_LOB0000077480C00002$$ 4.6 LOBSEGMENT T465 5 TABLE T2052 8.3 TABLE T2115 12.4 TABLE T2444 13.4 TABLE T2179 14.8 TABLE T2192 21.8 TABLE SYS_LOB0000077549C00015$$ 182 LOBSEGMENT <=== (related to table T2192) SYS_LOB0000077549C00016$$ 184.4 LOBSEGMENT <=== (related to table T2192) 30 rows selected.
So, in the light of the foregoing, I noticed in the first query that contains the table T2192 21.8 G of LOBS and that C456 and C459 of same table columns (181,7 + 183.9) total = 365.6 G.SQL> select table_name, column_name, segment_name 2 from dba_lobs 3 where segment_name in ( 4 select segment_name from dba_segments where owner = 'ARADMIN' 5 having round(sum(bytes)/1024/1024/1024,1) > 1 6 group by segment_name 7 ) 8 / TABLE_NAME COLUMN_NAME SEGMENT_NAME -------------------------------- -------------------------------- -------------------------------- B1947C536880923 C536880923 SYS_LOB0000077310C00002$$ T2051 C536870998 SYS_LOB0000077426C00041$$ T2052 C536870987 SYS_LOB0000077440C00063$$ T2115 C536870913 SYS_LOB0000077463C00009$$ B2125C536880912 C536880912 SYS_LOB0000077480C00002$$ B2125C536880913 C536880913 SYS_LOB0000077483C00002$$ T2179 C536870936 SYS_LOB0000077517C00027$$ T2192 C456 SYS_LOB0000077549C00015$$ <==== T2192 C459 SYS_LOB0000077549C00016$$ <==== T2444 C536870936 SYS_LOB0000210343C00029$$ T1990 C536870937 SYS_LOB0000250271C00026$$ 11 rows selected.
First question is how can the table be only 21.8 G, and business segments the columns of the table of 365,6 G of the Lobs?
It seems that some type LOB data should be external, while others are part of the table.
Then, I wonder if a line is removed from the table, would be the LOB associated with this line that is referenced by columns C456 and C459 also deleted.
Discuss with our senior developer, he said, the table is purged of lines longer than 6 months, but my question is if the LOB is actually purged with the lines.
Any ideas?
Published by: 974632 on December 27, 2012 08:05The space occupied by business segments is not reported as part of space for table segment, welll, don't not all still. What is reported as part of the size of the table is the size of the lob index (a pointer to the segment of the 'real' job, 16 bytes if I remember correctly) and the LOB that are stored online in the table (those under 4 k). To get the "full" size of the table, you must add the space occupied by the ant lob columns.
If you delete a line, then the space occupied in the business segments will be defined as free space inside the existing segment. He will thereafter (there are a number of round redo and undo that are managed differently from the lobs), be used by the data for the other lines, but it might take some time.
There are a number of documents on My Oracle support in this regard, many are related to:
Troubleshooting Guide (TSG) - large objects (LOBs) [ID 846562.1]
John
-
Hello experts
There are some queries to view and recover the unused space of LOV in non-spatail tablein this link http://www.idevelopment.info/data/Oracle/DBA_tips/LOBs/LOBS_85.shtml
We have geometry with sdo_geometry tables and their associates LOB sdo_ordinates and Ansdo_elem_info.
Can I apply these requests mentioned in the link to the spatial tables to recover unused space? If yes how?
Thank you
Can you please try something similar to the following?
ALTER table table_name modify varray geom_column. SDO_ORDINATES (retraction space);
ALTER table table_name modify varray geom_column. Ansdo_elem_info (space of retraction);
or
ALTER table table_name modify varray geom_column. SDO_ORDINATES (narrowing space) modify varray geom_column. Ansdo_elem_info (space of retraction);
-
ORA-02327: cannot create indexes on the LOB data type expression
DB 10.2.0.4
DB 5.2
I am not able to move lobindex and lobsegment. While try occurs below error
SQL > ALTER INDEX dev. "" SYS_IL0000718396C00002$ $"REBUILD THE TABLESPACE DATA_TEMP;
ALTER INDEX dev. "" SYS_IL0000718396C00002$ $"REBUILD TABLESPACE DATA_TEMP
*
ERROR on line 1:
ORA-02327: cannot create indexes on the LOB data type expression
Thank you
Published by: user13382934 on October 16, 2012 14:05correct syntax is
ALTER TABLE.MOVE the LOB (
)
STORE AS (tablespace); Where
is a name of the / columns LOB, not the name of the segment. Moving type LOB data will move its indexes as well. -
I can't create a table with the LOB data type.
Do I need an additional parameter in sys.odbc.ini?
My dsn information is
[tpch]
Driver=/home/TimesTen/TimesTen/tt1121/lib/libtten.so
Data store = / home/timesten/TimesTen/tt1121/tpch/tpch
LogDir = / home/timesten/TimesTen/tt1121/tpch/logs
PermSize = 4096
TempSize = 2048
PLSQL = 1
LOB = 1
DatabaseCharacterSet = US7ASCII
-bash-4, $ 1 ttdaemonadmin-version
TimesTen Release 11.2.1.2.0
There is no LOB = 1 attribute; Please, delete it from your definition ODBC to avoid potential errors.
LOB support has been added in the major version 11.2.2 TimesTen. So you use 11.2.1 LOB is not available. If you want to use LOB data, then you must upgrade to 11.2.2.
Chris
-
Where to define new custom data types?
Hello
In the past (TS 3.5) I created our own range of custom type file that has been used
to store the new data types and then referred the case to other colleagues. The file would be
stored in the ...\Program Files\...\User area.
My question arose because we now use the TS 4.1/4.2, which is no longer a separate
Directories of NOR and user in \Program Files.
Because now I want to change an existing custom data type, I find that our range of custom type has
fallen by the way side, forgotten.
Even though I can see the custom data definitions within the sequence of type files that use
custom data types, which means that I can edit them locally, I intend to return to
a range of custom type, i.e. comprehensive definition.
What is the relationship between the definitions in a range of custom type and custom data
definitions of type of a file of sequence?
When a palette to update file a file of sequence?, who takes over in the event of conflicts?
is a really necessary sequence files palette file if they are separated by using the same custom data type
can update the other? What is good practice when defining custom types of data?
Thank you
Gary.
Hey guys,.
This is a very interesting thread, and I've got everything right, in heart, with advice given so far. I just wanted to offer some additional tips on the conflicts of kind - with more response, the initial question concerning the definition of what is priority in case of conflict.
It is important to note that TestStand uses type names and version numbers to identify the different types. It is also important to note that when you use a definition of type of customers within a sequence, the sequence (.seq) file containing the sequence will keep a copy of the type definition. This greatly facilitates distributed sequence files. However, it also opens the door to potential conflicts type.
TestStand allows only one type of unique name to be loaded into memory at any given time, so that it uses the number of versions of the type to try to resolve these conflicts automatically. For example, TestStand can be configured to load any type is the largest version number (note that this can be changed via the tab Preferences dialog box Options of Station).
All this information and more are found in the following tutorials...
Conflicts and TestStand Type Versioning
How to make a Type of custom step?
Thanks for your time. I hope this has been helpful!
-
Uses: Oracle9i Standard Edition, ODAC 11.2.0.4, Oracle Instant Client 11.2.0.4
I am to convert the Microsoft Oracle for Oracle ODP.NET provider of my projects
* Error *.
Cannot be cast to type 'Oracle.DataAccess.Types.OracleDecimal' object type 'System.IConvertible '. Could not store < 145982 > in the RECEIPTID column. Type is Int64.
This error is poped up when my stored Proc has a return parameter. The return is a numeric data type in Oracle. I had no problem when using Oracle Provider from Microsoft.
Initially, I had my command setting insert the value
System.Data.OracleClient.OracleType.Number
for the column RECEIPTID; After ODP.NET, I converted to
Oracle.DataAccess.Client.OraclDbType.Double
Then the error throwed upward. Implement
Oracle.DataAccess.Client.OracleDbType.Int64
as suggested by the error, but nothing seems to prevent the error a way.
I am running using the OracleDataAdapter.
The OracleCommand object initialization *.
public class ReceiptAccountBusinessTable : EquityBroker32.APP_DATA.config.GenericBusinessTable { OracleCommand[] nCommandCollection; public ReceiptAccountBusinessTable() { this.Tablename = "JKSBSCHEMA.RECEIPTACCOUNT"; this.KeyDBField = "RECEIPTID"; this.DefaultSql = string.Format("SELECT * FROM {0} WHERE {1}= (SELECT MAX({1}) FROM {0} )", this.Tablename, this.KeyDBField); //select command this.nCommandCollection = new Oracle.DataAccess.Client.OracleCommand[4]; this.nCommandCollection[0] = new OracleCommand(); this.nCommandCollection[0].Connection = this.DataConnection; this.nCommandCollection[0].CommandText = this.DefaultSql; this.nCommandCollection[0].CommandType = System.Data.CommandType.Text; //update command this.nCommandCollection[1] = new OracleCommand(); this.nCommandCollection[1].Connection = this.DataConnection; this.nCommandCollection[1].CommandText = "JKSBSCHEMA.RECEIPTACCOUNT_ADDCHG_PROC"; this.nCommandCollection[1].BindByName = true; this.nCommandCollection[1].CommandType = System.Data.CommandType.StoredProcedure; this.nCommandCollection[1].Parameters.Add(new Oracle.DataAccess.Client.OracleParameter("PRECEIPTID", Oracle.DataAccess.Client.OracleDbType.Int64, 10, System.Data.ParameterDirection.InputOutput, false, 10, 0, "RECEIPTID", System.Data.DataRowVersion.Current, null)); this.nCommandCollection[1].Parameters.Add(new Oracle.DataAccess.Client.OracleParameter("PACCOUNTID_NEW", Oracle.DataAccess.Client.OracleDbType.Char, 13, System.Data.ParameterDirection.Input, false, 0, 0, "CLIENTACCOUNTID", System.Data.DataRowVersion.Current, null)); this.nCommandCollection[1].Parameters.Add(new Oracle.DataAccess.Client.OracleParameter("PACCOUNTID_OLD", Oracle.DataAccess.Client.OracleDbType.Char, 13, System.Data.ParameterDirection.Input, false, 0, 0, "CLIENTACCOUNTID", System.Data.DataRowVersion.Original, null)); this.nCommandCollection[1].Parameters.Add(new Oracle.DataAccess.Client.OracleParameter("PXSACTVALUE", Oracle.DataAccess.Client.OracleDbType.Double, 22, System.Data.ParameterDirection.Input, false, 0, 0, "XSACTVALUE", System.Data.DataRowVersion.Current, null)); this.nCommandCollection[1].Parameters.Add(new Oracle.DataAccess.Client.OracleParameter("PXSACTDESC", Oracle.DataAccess.Client.OracleDbType.Varchar2, 255, System.Data.ParameterDirection.Input, false, 0, 0, "XSACTDESC", System.Data.DataRowVersion.Current, null)); this.nCommandCollection[1].Parameters.Add(new Oracle.DataAccess.Client.OracleParameter("PPAGEBREAK", Oracle.DataAccess.Client.OracleDbType.Int32, 22, System.Data.ParameterDirection.Input, false, 0, 0, "PAGEBREAK", System.Data.DataRowVersion.Current, null)); //insert command this.nCommandCollection[2] = new OracleCommand(); this.nCommandCollection[2].Connection = this.DataConnection; this.nCommandCollection[2].CommandText = "JKSBSCHEMA.RECEIPTACCOUNT_ADDCHG_PROC"; this.nCommandCollection[2].CommandType = System.Data.CommandType.StoredProcedure; this.nCommandCollection[2].BindByName = true; this.nCommandCollection[2].Parameters.Add(new Oracle.DataAccess.Client.OracleParameter("PRECEIPTID", Oracle.DataAccess.Client.OracleDbType.Int64, 10, System.Data.ParameterDirection.InputOutput, false, 10, 0, "RECEIPTID", System.Data.DataRowVersion.Proposed, null)); this.nCommandCollection[2].Parameters.Add(new Oracle.DataAccess.Client.OracleParameter("PACCOUNTID_NEW", Oracle.DataAccess.Client.OracleDbType.Char, 13, System.Data.ParameterDirection.Input, false, 0, 0, "CLIENTACCOUNTID", System.Data.DataRowVersion.Current, null)); this.nCommandCollection[2].Parameters.Add(new Oracle.DataAccess.Client.OracleParameter("PACCOUNTID_OLD", Oracle.DataAccess.Client.OracleDbType.Char, 13, System.Data.ParameterDirection.Input, false, 0, 0, "CLIENTACCOUNTID", System.Data.DataRowVersion.Original, null)); this.nCommandCollection[2].Parameters.Add(new Oracle.DataAccess.Client.OracleParameter("PXSACTVALUE", Oracle.DataAccess.Client.OracleDbType.Double, 22, System.Data.ParameterDirection.Input, false, 0, 0, "XSACTVALUE", System.Data.DataRowVersion.Current, null)); this.nCommandCollection[2].Parameters.Add(new Oracle.DataAccess.Client.OracleParameter("PXSACTDESC", Oracle.DataAccess.Client.OracleDbType.Varchar2, 255, System.Data.ParameterDirection.Input, false, 0, 0, "XSACTDESC", System.Data.DataRowVersion.Current, null)); this.nCommandCollection[2].Parameters.Add(new Oracle.DataAccess.Client.OracleParameter("PPAGEBREAK", Oracle.DataAccess.Client.OracleDbType.Int32, 22, System.Data.ParameterDirection.Input, false, 0, 0, "PAGEBREAK", System.Data.DataRowVersion.Current, null)); //delete command this.nCommandCollection[3] = new OracleCommand(); this.nCommandCollection[3].Connection = this.DataConnection; this.nCommandCollection[3].CommandText = "DELETE FROM " + this.Tablename + " WHERE RECEIPTID=:p1 AND ACCOUNTID=RPAD(:p2,15)"; this.nCommandCollection[3].CommandType = System.Data.CommandType.Text; this.nCommandCollection[3].Parameters.Add(new Oracle.DataAccess.Client.OracleParameter("p1", Oracle.DataAccess.Client.OracleDbType.Int64, 22, System.Data.ParameterDirection.Input, false, 0, 0, "RECEIPTID", System.Data.DataRowVersion.Current, null)); this.nCommandCollection[3].Parameters.Add(new Oracle.DataAccess.Client.OracleParameter("p2", Oracle.DataAccess.Client.OracleDbType.Char, 15, System.Data.ParameterDirection.Input, false, 0, 0, "ACCOUNTID", System.Data.DataRowVersion.Current, null)); this.CommandCollection = this.nCommandCollection; } }* The glance of the stored Proc *.
PROCEDURE RECEIPT_ADDCHG_PROC ( preceiptid IN OUT NUMBER, preceiptmode IN CHAR, pxsactdate IN DATE, pxsactvalue IN NUMBER, pbankaccountid IN VARCHAR2, pchequeno IN VARCHAR2, pclientbankid IN VARCHAR2, pdescription IN VARCHAR2, pstatementno in varchar2 default null, pcustodypayment IN NUMBER) is nrow number; begin --check if posted SELECT COUNT(*) INTO nrow from accountledger WHERE receiptid = preceiptid; IF nrow > 0 THEN RAISE_APPLICATION_ERROR(-20001, 'Record Already Posted'); END IF; UPDATE Receipt Set pmodeid = preceiptmode, xsactdate = pxsactdate, xsactvalue = pxsactvalue, bankaccountid = pbankaccountid, chequeno = pchequeno, clientbankid = pclientbankid, description = pdescription, statementno = pstatementno, custodypayment = pcustodypayment WHERE receiptid = preceiptid; IF (SQL%NOTFOUND) THEN INSERT INTO RECEIPT(receiptid, pmodeid, xsactdate, xsactvalue, bankaccountid, chequeno, clientbankid, description, statementno, custodypayment ) VALUES(preceiptid, preceiptmode, pxsactdate, pxsactvalue, pbankaccountid, pchequeno, pclientbankid, pdescription, pstatementno, pcustodypayment ) RETURNING receiptid into (preceiptid); END IF; end;
* It's how I update the Table *.
public void UpdateData(DataTable tbl, Oracle.DataAccess.Client.OracleTransaction oTran) { try { SetTransaction(oTran); this._adapter.Update(tbl); //this was added on 03-Aug-2011 tbl.AcceptChanges(); } //handling concurrency problem catch (DBConcurrencyException dbcx) { MessageBox.Show(CreateMessage(dbcx.Row), "Concurrency Exception", MessageBoxButtons.YesNo); } catch (OracleException ex) { throw ex; } catch (Exception ex) { throw ex; } }* The said Stack Trace *.
at System.Data.Common.DbDataAdapter.UpdatedRowStatusErrors (Int32 commandCount RowUpdatedEventArgs rowUpdatedEvent, BatchCommandInfo [batchCommands])
at System.Data.Common.DbDataAdapter.UpdatedRowStatus (Int32 commandCount RowUpdatedEventArgs rowUpdatedEvent, BatchCommandInfo [batchCommands])
to System.Data.Common.DbDataAdapter.Update (DataRow [] dataRows, DataTableMapping tableMapping)
to Oracle.DataAccess.Client.OracleDataAdapter.Update (DataRow [] dataRows, DataTableMapping tableMapping)
at System.Data.Common.DbDataAdapter.UpdateFromDataTable (DataTable dataTable, DataTableMapping tableMapping)
at System.Data.Common.DbDataAdapter.Update (DataTable dataTable)
at EquityBroker32.APP_DATA.config.GenericBusinessTable.UpdateData (DataTable tbl, OracleTransaction oTran) in d:\REVERSION\Latest\APP_DATA\config\GenericBusinessTable.cs:line 293This is found!
I add the following code to my existing and it worked!
this.nCommandCollection[1].Parameters.Add(new Oracle.DataAccess.Client.OracleParameter("PRECEIPTID", Oracle.DataAccess.Client.OracleDbType.Double, 10, System.Data.ParameterDirection.InputOutput, false, 10, 0, "RECEIPTID", System.Data.DataRowVersion.Current, null)); this.nCommandCollection[1].Parameters["PRECEIPTID"].DbType = DbType.Decimal;Thanks to this site and the person who answered: .Net technology: find a sequence of a S-P with typed DataSet problem - dotnet - tech.todaysummary.com
-
Dynamic SQL with in bulk in the record type
Oracle 10.2 g
I received this Tom
[http://asktom.oracle.com/pls/apex/f?p=100:11:0:NO:]
I'm able to do this without dynamic SQL, but my requirement is to do it in dynamic SQL
There is a work around in 11g, but can we do something in 10g?create table t1 ( x int, y int ); insert into t1 select rownum, rownum+1 from all_users where rownum <= 5; create table t2 ( x int, y int, z int ); declare type array is table of t1%rowtype; l_data array; begin select * bulk collect into l_data from t1; forall i in 1 .. l_data.count execute immediate 'insert into (select x, y from t2) values :x' using l_data(i); end; Error at line 1 ORA-06550: line 9, column 90: PLS-00457: expressions have to be of SQL types ORA-06550: line 9, column 20: PL/SQL: Statement ignored
Thank you
HESH.HESH wrote:
but following does not.
declare type array is table of t1%rowtype; l_data array; begin select * bulk collect into l_data from t1; forall i in 1 .. l_data.count execute immediate 'insert into (select x, y from t2) values :x' using l_data(i); end;I want just a dynamic SQL code for the insert with FORALL statement would adopt as well as collections.
Doesn't make much sense.
Extract you the data from the SQL engine in the table of the record type. If the output data that cursor SQL must be read in the SQL engine and copied into the memory of PL/SQL engine.
Then, you send that VERY SAME DATA back to the SQL engine to be used by a SQL insert cursor.
Where is the logic behind the extraction of data from SQL in a PL/SQL table structure and then push this same structure table on the SQL engine database? What is the purpose to send data on a detour of underperforming and non-scalale through the PL/SQL engine?
You have any justification (technical or functional wise) to back up this absurd approach?
Why this can be achieved using a single SQL cursor that does both the choice (extraction) and (in bulk) insertion - using the plain old INSERT... SELECT structure?
And if the insert is variable, then what? Create a dynamic INSERT... SELECT cursor and execute it (using bind values). This simple... Right?
-
SQL Developer 3.0 EA 1 Advanced support for the type of data in the unit tests
Hello
According to SQL Developer 3.0 EA 1 New Feature List (http://www.oracle.com/technetwork/developer-tools/sql-developer/rel3-featurelist-ea1-166831.html#ut)
the EA 1 version 3.0 supported advanced data types in the unit tests. This means support the ANYDATA and RECORD types? I installed version 3.0 of EA 1 to CentOS linux, but it seems that those advanced types are not supported yet. When I try to add a unit test for a function that returns the record type, I get quick error: "the return of PL/SQL RECORD type is not supported. Cannot test (function_name) because 1 arguments have not supported for types". Also if I try to unit test procedure which parameter is of type ANYDATA, I get quick error: "the type UNDEFINED P_DATA argument is not supported. Cannot test (procedure_name) because 1 arguments have not supported for types". Are those advanced data types supported in SQL Developer 3.0 FINAL version?
Thank you
Miikka L
Edited by: user12844253 the 27.10.2010 16:31Hi Mikka,
Only simple PL/SQL records are currently supported i.e. those containing no optional or repeating components and where all components are themselves supported. This restriction is in place, due to the fact that we are using JDBC as a parameter mechanism which does not directly support the passage the PL/SQL record type.
ANYTYPE and ANYDATA are currently not supported because they have a dynamic value type and must be set programmatically. In the future, it would be possible to support these via the dynamic value and and validation features.
This will remain for the final version.
Kind regards
Richard -
Hello experts.
create type emp2_obj is object
(
objno number,
objname varchar2(20),
objdept number
);
create type emp2_objarr is table of emp2_obj;and
type emp2_rec is record
(
recno number,
recname varchar2(20),
recdept number
);
create type emp2_recarr is table ofemp2_rec;Objects and types of records are created similar and have the same similar object. Is it only advisable to use registration type if you use a collection in PL/SQL type... Please advice
user13328581 wrote:
the only reason why I ask is because he asked during an interview
-Objects and the types of records are created similar and have the same similar object. Is it only advisable to use registration type if you use a collection in PL/SQL type... Please advice
Not at all. You can use record types used with collection types outside. They are a group concept. If you want to group a set of values together (including fields), you can use a record. It may be convenient. You can send documents around proc and functions.
Be aware that the record types are limited to procedures, functions, anonymous blocks and packages (i.e. PL/SQL), while the types of objects are stored in the dictionary of data as a separate Oracle objects (they can also have methods, the records may not). Object types can be used in SQL, as said sol.beach types of records (outside of the intelligent pipeline situations) is for PL/SQL.
-
HelloI have a requirement of the company, where I need to return a record type (OUT parameter) for environment call based on the given input value.
Suppose that if the value is correct and corresponding record is found in the table then the return values for this key entry. If matching record is found, then return the exception to the calling environment.
To do this, I created an example of test table and populated records.
create table plch_test(dept_id number,dept_name varchar2(50),cost_centre number); insert into plch_test values(10,'SALES',1010); insert into plch_test values(20,'FINANCE',2010); insert into plch_test values(30,'MKTG',3010); SQL> select * from plch_test; DEPT_ID DEPT_NAME COST_CENTRE ---------- -------------------------------------------------- ----------- 10 SALES 1010 20 FINANCE 2010 30 MKTG 3010
I wrote a simple block and gave a valid key dept_id (10 in this case) to display costcentre for this dept_id and dept_name I said tow types of records, one for valid record and another exception
SQL> DECLARE 2 TYPE rec_dept IS RECORD(dept_name varchar2(50),cc number); 3 l_rec_dept rec_dept; 4 TYPE rec_exception IS RECORD(err_code number,error_message varchar2(300)); 5 l_rec_exception rec_exception; 6 BEGIN 7 SELECT dept_name,cost_centre 8 INTO l_rec_dept 9 FROM plch_test 10 where dept_id=10; 11 dbms_output.put_line('DEPT_NAME'||' '||l_rec_dept.dept_name||' '||'COSTCENTRE'||' '||l_rec_dept.cc); 12 EXCEPTION WHEN NO_DATA_FOUND THEN 13 l_rec_exception.err_code:=sqlcode; 14 l_rec_exception.error_message:=sqlerrm; 15 dbms_output.put_line(l_rec_exception.err_code||' '||l_rec_exception.error_message); 16 END; 17 . SQL> / DEPT_NAME SALES COSTCENTRE 1010 PL/SQL procedure successfully completed. SQL>Now for invalid dept_id and expose the message by using exception record type I stated.
SQL> ed Wrote file afiedt.buf 1 DECLARE 2 TYPE rec_dept IS RECORD(dept_name varchar2(50),cc number); 3 l_rec_dept rec_dept; 4 TYPE rec_exception IS RECORD(err_code number,error_message varchar2(300)); 5 l_rec_exception rec_exception; 6 BEGIN 7 SELECT dept_name,cost_centre 8 INTO l_rec_dept 9 FROM plch_test 10 where dept_id=40; --Invalid --data is not present 11 dbms_output.put_line('DEPT_NAME'||' '||l_rec_dept.dept_name||' '||'COSTCENTRE'||' '||l_rec_dept.cc); 12 EXCEPTION WHEN NO_DATA_FOUND THEN 13 l_rec_exception.err_code:=sqlcode; 14 l_rec_exception.error_message:=sqlerrm; 15 dbms_output.put_line(l_rec_exception.err_code||' '||l_rec_exception.error_message); 16* END; SQL> / 100 ORA-01403: no data found PL/SQL procedure successfully completed.Now as you can see I need to include this point in a procedure with an input parameter and output must be a record types which will return
rec_dept if it becomes a key input valid or an exception if she meets a key not valid.
CREATE PROCEDURE test_prc IS(p_in_dept_id IN plch_test.dept_id,p_output ?????? DECLARE TYPE rec_dept IS RECORD(dept_name varchar2(50),cc number); l_rec_dept rec_dept; TYPE rec_exception IS RECORD(err_code number,error_message varchar2(300)); l_rec_exception rec_exception; BEGIN BEGIN SELECT dept_name,cost_centre INTO l_rec_dept FROM plch_test where dept_id=p_ind_dept_id; RETURN l_rec_dept; EXCEPTION WHEN NO_DATA_FOUND THEN l_rec_exception.err_code:=sqlcode; l_rec_exception.error_message:=sqlerrm; RETURN l_rec_exception; END; dbms_output.put_line('DEPT_NAME'||' '||l_rec_dept.dept_name||' '||'COSTCENTRE'||' '||l_rec_dept.cc); END;Hope that the explanation above help in imposes the requirement
Kind regards
Claudy kotekal
Return a record which can mean two things is complicated; I'm not an experienced myself pl/sql developer, but this looks like a craft.
The idea of exceptions under Sir Thomas of Kyte, is that any treatment must be stopped; You should RAISE an exception to the appellant so that he can figure out what to do with it. What you are saying, this is an exception, but is not a little, cos it's okay, I'll just keep but I will go back to the appellant in any way, but the appellant shall include this registration type is - would it be a record representing a row of the table, or it might be an exception... yuck.
(a) is it really an exception
(b) what do you do with it? You he could log into a table, you could write to a file, you can display an error message on the screen
But really, it's weird to want to pass an exception as return value.
These are all considerations of design, not really anything to do with the pl/sql language in itself.
But hard, if you send a record type a successful being found, registration-based stick to it and don't use it to return a record; do not try to do double duty with her flipping something else. Just save the message put in a table, or print it to the console, or what you want to do with; but as I said, the most important decision is, is this really an exception. And is based on the data model and the expectations of cleanliness of the data etc.
Think about how you call built-in functions. If you send garbage to a built-in function it does not return successfully, leaving you to figure out whether he succeeded or not by inspecting the return value; It goes kaboom, something bad happened. That's what your function should do if something bad happens, that is to say, if you get an exception, it should probably go kaboom.
-
PLS-00362: Invalid cursor return type; 'NUMBER' must be a record type
Hello
Having a little trouble with the following code example provided to http://www.dba-oracle.com/plsql/t_plsql_cursor_variables.htm:
In the code above, SELECT COUNT (*)... returns a NUMBER. I know it's an aggregation function, but it returns a single value.1 DECLARE 2 TYPE t_ref_cursor IS REF CURSOR RETURN NUMBER; 3 c_cursor t_ref_cursor; 4 l_row NUMBER; 5 BEGIN 6 DBMS_OUTPUT.put_line('Strongly typed REF CURSOR using SCALAR type. Expect an error!'); 7 OPEN c_cursor FOR 8 SELECT COUNT(*) cnt 9 FROM cursor_variable_test; 10 LOOP 11 FETCH c_cursor 12 INTO l_row; 13 EXIT WHEN c_cursor%NOTFOUND; 14 DBMS_OUTPUT.put_line(l_row); 15 END LOOP; 16 CLOSE c_cursor; 17* END; 18 / TYPE t_ref_cursor IS REF CURSOR RETURN NUMBER; * ERROR at line 2: ORA-06550: line 2, column 24: PLS-00362: invalid cursor return type; 'NUMBER' must be a record type ORA-06550: line 2, column 3: PL/SQL: Item ignored
Why can't return a value in a column of a row in a NUMBER?
How can I change the SQL code so that I can do this?
Furthermore, I wonder about the use of FETCH with a count (*)... FETCH is supposed to fetch the next row... How it works when you select an aggregate as County?
Thank you very much
Jason>
TYPE t_ref_cursor IS REF CURSOR RETURN NUMBER;
*
ERROR on line 2:
ORA-06550: line 2, column 24:
PLS-00362: Invalid cursor return type; 'NUMBER' must be a record type
ORA-06550: line 2, column 3:
PL/SQL: Ignored ElementIn the code above, SELECT COUNT (*)... returns a NUMBER. I know it's an aggregation function, but it returns a single value.
Why can't return a value in a column of a row in a NUMBER?
How can I change the SQL code so that I can do this?
>
The exception is in line 2: your cursor statement. And the answer is in the text that you access
>
The return value of a strongly typed REF CURSOR must be a folder that can be defined using % TYPE % ROWTYPE attributes or record structure.
>
You said the CURSOR to return a NUMBER. And as the text says, he must be a 'record '.
>
Furthermore, I wonder about the use of FETCH with a count (*)... FETCH is supposed to fetch the next row... How it works when you select an aggregate as County?
>
As you said already FETCH retrieves the next line, if any. A query is a request is a request. It returns a result set. A query that uses aggregates returns a result set. A query that does not aggregate returns a result set.Your simple COUNT (*) SELECT query returns a result set that consists of a LINE and a line a ONE COLUMN of type NUMBER. Although there is only one column in the result set, what is returned is a RECORD or a LINE. That's why you have to report your data cursor return type a document using the % ROWTYPE or % TYPE attributes or a record structure.
-
Update a primary key of type char data
Hello
Not a guy to oracle technology. but am forced to work on a query that is a basic requirement.
I have a table that has a composite primary key values (6 columns).
Of these, I've updated 1 specific column, which is a char data type.
I tested after writing my logic as a query, threw it [the Unique constraint violated schema.tablename];
So I gave it a query simple base to test it yet, still get the same error.
Query such as:
Update table_name set Column1 = "100000" where column2 = 'BOMBAY '; (Column1 is the primary key of type char)
Even if it did not work.
Now a doubt arises, a column is a primary key of type Char can be modified? or is it not possible at all to do?
Note: I made sure only am not updated with a duplicate value;
Thank you.
Published by: 966353 on October 18, 2012 12:01While you never need to update a primary key (regardless of the type of data), it is possible to do. If you get an error ORA-00001 indicating that a unique constraint has been violated and the unique constraint that is violated is associated with the primary key of the table (I'm guessing that this is the case according to your description of the error), implying strongly that the update would result in duplicate rows. Your UPDATE statement updates more than 1 row? If so, how do you determine when the update is complete there will be no duplicate lines?
Justin
-
Error using BULK collect with RECORD TYPE
Hello
I wrote a simple procedure to declare a record type & then by a variable of type NESTED table.
I then selects the data using COLLECT in BULK & trying to access it via a LOOP... We get an ERROR.
------------------------------------------------------------------------------------------------------------------------------------------------------
CREATE OR REPLACE PROCEDURE sp_test_bulkcollect
IS
TYPE rec_type () IS RENDERING
emp_id VARCHAR2 (20).
level_id NUMBER
);
TYPE v_rec_type IS TABLE OF THE rec_type;
BEGIN
SELECT employe_id, level_id
LOOSE COLLECTION v_rec_type
OF portfolio_exec_level_mapping
WHERE portfolio_execp_id = 2851852;
FOR indx IN v_rec_type. FIRST... v_rec_type. LAST
LOOP
dbms_output.put_line ('Emp-' | v_rec_type.emp_id (indx) |) » '|| v_rec_type.level_id (indx));
END LOOP;
END;
-----------------------------------------------------------------------------------------------------------------------------------
Here is the ERROR I get...
-Errors of compilation for the PROCEDURE DOMRATBDTESTUSER. SP_TEST_BULKCOLLECT
Error: PLS-00321: expression "V_REC_TYPE" is not appropriate for the left side of an assignment statement
Online: 15
Text: IN portfolio_exec_level_mapping
Error: PL/SQL: ORA-00904: invalid identifier
Online: 16
Text: WHERE portfolio_execp_id = 2851852;
Error: PL/SQL: statement ignored
Line: 14
Text: COLLECT LOOSE v_rec_type
Error: PLS-00302: component 'FIRST' must be declared
Online: 19
Text: LOOP
Error: PL/SQL: statement ignored
Online: 19
Text: LOOP
------------------------------------------------------------------------------------------------
Help PLZ.and with a complete code example:
SQL> CREATE OR REPLACE PROCEDURE sp_test_bulkcollect 2 IS 3 TYPE rec_type IS RECORD ( 4 emp_id VARCHAR2(20), 5 level_id NUMBER 6 ); 7 TYPE v_rec_type IS TABLE OF rec_type; 8 v v_rec_type; 9 BEGIN 10 SELECT empno, sal 11 BULK COLLECT INTO v 12 FROM emp 13 WHERE empno = 7876; 14 FOR indx IN v.FIRST..v.LAST 15 LOOP 16 dbms_output.put_line('Emp -- '||v(indx).emp_id||' '||v(indx).level_id); 17 END LOOP; 18 END; 19 / Procedure created. SQL> SQL> show error No errors. SQL> SQL> begin 2 sp_test_bulkcollect; 3 end; 4 / Emp -- 7876 1100 PL/SQL procedure successfully completed. -
Foreground of Oracle process to access data under Linux files
As far as I know, processes user in an Oracle database running the application code, but the processes of database running Oracle database server code (server process analyze and execute the SQL statements issued through the app, read data of data files blocks and return the results to the application.) Background treats the archive logs, update the headers of all the data files to save the details of the control point, written content of the buffers of data files, runs the recovery if necessary, etc.)
However, I could see the oracle user processes to access data files in my Oracle database. When I look for the processes to access data belonging to the tablespace files "recprov", I get the following:
[root@ymir ~] # lsof | grep recprov
Oracle Oracle 465 11u REG 8.18 6291464192 /ora2/oradata/essepr3/recprov_tb_01.dbf 48496658
Oracle Oracle 465 13u REG 8.18 4823457792 /ora2/oradata/essepr3/recprov_tb_02.dbf 48496659
Oracle Oracle 465 15u REG 8.18 4823457792 /ora2/oradata/essepr3/recprov_tb_02.dbf 48496659
Oracle Oracle 964 11u REG 8.18 4823457792 /ora2/oradata/essepr3/recprov_tb_02.dbf 48496659
Oracle Oracle 964 13u REG 8.18 4823457792 /ora2/oradata/essepr3/recprov_tb_02.dbf 48496659
Oracle Oracle 964 14u REG 8.18 6291464192 /ora2/oradata/essepr3/recprov_tb_01.dbf 48496658
Oracle Oracle 13364 14u REG 8.18 4823457792 /ora2/oradata/essepr3/recprov_tb_02.dbf 48496659
Oracle Oracle 13364 76u REG 8.18 6291464192 /ora2/oradata/essepr3/recprov_tb_01.dbf 48496658
Oracle Oracle 16445 17u REG 8.18 4823457792 /ora2/oradata/essepr3/recprov_tb_02.dbf 48496659
Oracle Oracle 16445 18u REG 8.18 6291464192 /ora2/oradata/essepr3/recprov_tb_01.dbf 48496658
Oracle Oracle 20522 REG 8.18 4823457792 /ora2/oradata/essepr3/recprov_tb_02.dbf 48496659 82uW
Oracle Oracle 20522 REG 8.18 6291464192 /ora2/oradata/essepr3/recprov_tb_01.dbf 48496658 122uW
Oracle Oracle 20532 44u REG 8.18 4823457792 /ora2/oradata/essepr3/recprov_tb_02.dbf 48496659
Oracle Oracle 20532 75u REG 8.18 6291464192 /ora2/oradata/essepr3/recprov_tb_01.dbf 48496658
Oracle Oracle 20534 17u REG 8.18 4823457792 /ora2/oradata/essepr3/recprov_tb_02.dbf 48496659
Oracle Oracle 20534 79u REG 8.18 6291464192 /ora2/oradata/essepr3/recprov_tb_01.dbf 48496658
As you can see, there are 7 oracle processes that access data files: 4 processes user and 3 background processes:
[root@ymir ~] # ps - ef
...
Oracle 465 1 Jan27 0? 00:00:23 oracleessepr3 (LOCAL = NO)
Oracle 964 1 Jan27 0? 00:00:25 oracleessepr3 (LOCAL = NO)
Oracle 13364 1 0 Jan13? 02:45:02 oracleessepr3 (LOCAL = NO)
Oracle 16445 1 0 Jan21? 00:00:05 oracleessepr3 (LOCAL = NO)
20522 1 0 2009 Oracle? 00:09:59 ora_dbw0_essepr3
20532 1 0 2009 Oracle? 00:04:24 ora_ckpt_essepr3
20534 1 0 2009 Oracle? 00:04:10 ora_smon_essepr3
...
And I confirmed this information from the data base repository:
SQL > select p.SPID, SSE. TYPE, SSE. Username, SSE. MACHINE, SSE. PROGRAM
session $ v ESS, v$ process p
where p.SPID (465, 964, 13364, 16445, 20522, 20532, 20534) and SES. PADDR = p.ADDR;
USER NAME OF TYPE SPID MACHINE PROGRAM
----- ----------- --------------- --------------- ---------------------------------------------------------
20522 BACKGROUND dbserver (DBW0) oracle@dbserver
20532 BACKGROUND dbserver oracle@dbserver (CKPT)
20534 BACKGROUND dbserver oracle@dbserver (SMON)
13364 USER DBSNMP dbserver emagent@dbserver (TNS V1 - V3)
964 RECPROV USER appserver JDBC Thin Client
16445 USER SYSTEM userPC PlSqlDev.exe
465 RECPROV USER appserver JDBC Thin Client
* RECPROV is the name of the user of the application database.
My question is, why are there user processes to access data files? I think that it does not match the process architecture defined by Oracle. And most importantly, it does not mean a security leak in the database?
DB version: 10.2.0.3, dedicated
The Linux Version: Linux 2.6.18 - 53.el5xen #1 SMP Sat Nov 10 19:46:06 EDT 2007 x86_64 x86_64 x86_64 GNU/Linux
Thanks in advanceHello
There seems to be some confusion, even in the Oracle documentation, in terms of nomenclature.
Think of it this way. There is the client program that connects to Oracle, such as SQL * Plus, Toad, custom, etc. written application. This program runs on the client. This may or may not be on the database server.
Then there's the Oracle database server process. They run always on the database server. These processes are of two types, the USER and the background.
You can see:
SQL> select background,count(*) from v$process group by background; B COUNT(*) - ---------- 65 1 37Here are copies of the file binary oracle attach directly to the SGA and interact with database files. The binary oracle works with several different "personalities" based on the different purposes, he is currently serving. The type of background process, are divided into different specific methods for specific purposes.
Yet once, observe:SQL> select pname from v$process where background = 1; PNAME ----- PMON VKTM GEN0 DIAG DBRM PING PSP0 ACMS DIA0 LMON LMD0 PNAME ----- LMS0 LMS1 RMS0 LMHB MMAN DBW0 DBW1 LGWR CKPT SMON RECO PNAME ----- RBAL ASMB MMON MMNL MARK LCK0 RSMN GTX0 SMCO W000 RCBG PNAME ----- QMNC Q001 CJQ0 Q005 37 rows selected.Thus, each background process serves a different purpose. For example. DBW0 concerns only to write data from the buffer cache data files and maintaining all data structures associated with it. Each background process above serves a different purpose, I will not go into all. Generally, the background processes start up time of instance startup, even though some may be started on demand, at a certain time after the start of the proceeding.
Now, consider a dedicated server model. When you connect to the database by using SQL * Plus, for example, you run the binary "sqlplus". Assume that you are connected via SQL * Net. In this case, sqlplus connects to the listener, the listener creates an oracle server process, type 'USER '. This process always runs on the database server. It runs under the user 'oracle', and it provides your real and direct interface to the database. This user can interact with the database and read the database files directly.
Mainly, your server processes the USER will read the data from the cache buffers and, in the case of an absence of cache, will read the data in the data files and loading into the buffer cache. In addition, according to the code path, it can directly read or write to the datafiles, bypassing the buffer cache.
So, the bottom line, all server (USER and BOTTOM) processes, interact with the database and associated files. The code of client program (sqlplus, Toad, etc.) NEVER directly interacts with the database.
Hope that things cleared up.
-Mark
Maybe you are looking for
-
Install the LabVIEW DSP Module on Labview 2009 Professional
I have LabVIEW 2009 professional and I try to install the LabVIEW DSP Module on it. I have LabVIEW DSP Module 8.6, and I think that this is the final version. However, when I tried to install it, it says you should have LabVIEW 8.6 on your system. Wh
-
NASCAR Racing Sim does not work in Windows 7
I can't get my NASCAR Sim racing to open in 7. It worked fine under XP with Service Pack 2. Help please!
-
How to replace the graphics card Nvidia FX1500M in my Compac HP nw9440 (faulty after a fan failure). Works well with VGA HP chart but when FX1500M applied, it hangs at the blue screen. Suspect that the graphics card from Nvidia have suffered under e
-
Mail window problem - I can't send email
Windows mail problem (I can not send e-mails) Yaho tells me that it is a problem of Att. Att tells me that it is a windows vista problem. I can't get a clear answer form someone. Any thoughts? Thank you
-
BlackBerry Z30 only appealed to one please times
Hello Where can I put the Z30 take only one call at a time? When I'm busy speaking to someone I can hear other calls come in. I don't want to, if I'm busy, I want it to go to your voicemail.