Flush file before writing using UTL_FILE

Hi,
Version: 10g
I am using the UTL_FILE.PUT_LINE to write some data into a file.
Now my requirement is to flush this file before every write. Is there a function/procedure which does this?
Require help ASAP.
Thanks in advance.

Why flush before a write? The write buffer will be empty. And there will be nothing to flush.
You need to flush after a write in order to flush that write buffer to disk.
You also need to ask yourself why you want to do this? Write caches/lazy writes improve performance. The primary reason for flushing a cache is to ensure that the cache contents are on disk in case the process dies and its dirty unwritten cache is destroyed with it.
Do you expect your process to crash and die? If not, then what other reasons do you think warrants flushing the cache after each write?

Similar Messages

Reading all files on directory using "utl_file" package...

I need to read all files in directory via PL/SQL. I don't know
name files (are data dynamics create for automation system),
only I know your extensions.
Can I do this using the package "utl_file" or I need to create
program in another language (C, C++, for example)?
Any ideas...
Thanks.

Hi,
you can't do that with the UTL_FILE package (it can't retrieve
file names).
A very simple solution would be, if you created on OS-level a
file which contains the filenames of directory and then read this
file using UTL_FILE. With the information on all file names you
can enter a loop which opens and reads all files again using
UTL_FILE.
A more mundane solution could be to use the features on the iFS.
Cheers
Gerald

How can I import data from a csv file into databse using utl_file?

Hi,
I have two machines (os is windows and database is oracle 10g) that are not connected to each other and both are having the same database schema but data is all different.
Now on one machine, I want to take dump of all the tables into csv files. e.g. if my table name is test then the exported file is test.csv and if the table name is sample then csv file name is sample.csv and so on.
Now I want to import the data from these csv files into the tables on second machine. if I've 50 such csv files, then data should be written to 50 tables.
I am new to this. Could anyone please let me know how can I import data back into tables. i can't use sqlloader as I've to satisfy a few conditions while loading the data into tables. I am stuck and not able to proceed.
Please let me know how can I do this.
Thanks,
Shilpi

Why you want to export into .csv file.Why not export/import? What is your oracle version?
Read http://www.oracle-base.com/articles/10g/oracle-data-pump-10g.php
Regards
Biju

Renaming master files before import using metadata information

In order to be sure that none of my master files has the same names I am using the following naming format:
<camera model> YYYYMMDD HHMMSS
Example: Nikon D2X 20070218 213445
Until now I have used Photoshop CS2 in order to rename all my master files but it would be easier if is possible to automate the process. The date and time is taken from the metadata in each file. Which is the specific time the photo is taken and not just the "file creation date and time".
Has anyone found a tool/automation that could do this?

Karsten
Have a look at DIM:
http://www.alanlight.com/dim/Dim.htm
In order to be 'future proof' I import and rename all images into a Finder folder structure created by the above. I then import them into Aperture - so I have the same structure in the Finder and in Aperture. I keep al RAW images in a different structure from JPEG images.
DIM will do all you need
Brian

List of files (for processing using UTL_FILE)

We have an APEX (v2.2) application that needs to load data in from flat files. These files are in a directory on the server that is accessable to code in the UTL_FILE package and are simple csv and fixed format text files.
The problem is the file names change and the users want to see the filenames on screen in the APEX app so they can select which file to load in.
How can we get a list of files in a directory on the server to show on screen ? Once we have a filename, it can be processed by the UTL_FILE package code OK, we just need to be able to get the filenames !

On AskTom there is thread that will answer to you:
http://asktom.oracle.com/pls/asktom/f?p=100:11:0::::P11_QUESTION_ID:439619916584

FILE SIZE LIMIT WITH UTL_FILE?

Hello all,
is there any file size limit using utl_file on SUSE LINUX 5.3
(RDBMS 8.0.5)?
We have created a file with UTL_FILE.FOPEN, put some line into it
(UTL_FILE.PUT_LINE), and close the file (UTL_FILE.FCLOSE).
The file size is never greater than 32 kB.
On 7.3.4 the same procedure works fine with files larger than 1
MB.
Thx
R. Petuker
null

Robert Petuker (guest) wrote:
: Hello all,
: is there any file size limit using utl_file on SUSE LINUX 5.3
: (RDBMS 8.0.5)?
: We have created a file with UTL_FILE.FOPEN, put some line into
it
: (UTL_FILE.PUT_LINE), and close the file (UTL_FILE.FCLOSE).
: The file size is never greater than 32 kB.
: On 7.3.4 the same procedure works fine with files larger than 1
: MB.
: Thx
: R. Petuker
Robert,
There is no limit on the size of the file created using the
utl_file package. It could be the limitation of the shell
above which you are working on. You may verify the filesize
limit of the shell and try it again.
Maturi
null

Error while writing to file using UTL_FILE

I've created below mentioned directory object and also granted the user appdevnew the appropriate previleges under SYSTEM user. Before giving this command, i've manually created the folder in c drive and created myspool.log file manually. Is it needed or can be created programmatically?
CREATE DIRECTORY mydir AS 'C:\mydir';
GRANT READ,WRITE ON DIRECTORY MYDIR TO appdevnew;
Then logged into appdevnew and ran the block below. But i had been shown with error code ora-29280.
DECLARE
hnd UTL_FILE.FILE_TYPE;
BEGIN
hnd := UTL_FILE.FOPEN ('mydir', 'myspool.log', 'w',80);
UTL_FILE.PUT_LINE (hnd, 'Some text here..', TRUE);
UTL_FILE.FCLOSE (hnd);
END;
/

Hi,
It's a common problem, people think utl_file can see their PC's, whereas in fact the code runs on a server sometimes miles away and has no clue about your PC! You aren't the first and you won't be the last either.
As for creating a directory, I take it you mean to somehow issue 'mkdir' or the equivalent to create an OS level directory (or Folder, as this is Whinedoze) on the server? You have to be careful with terminology here, because people will think (and quite rightly to) that you are referring to directory objects which utl_file now uses.
There is no native way of doing this in PL/SQL. You can if you wish write a java stored procedure which then gives you this ability inside PL/SQL. However, I would question the need to actually dynamically create OS directories. Better do it just once for setup and be done. Less moving parts, less headaches.
And I don't know what you mean by the file extension at all. You opened a file with the extension .log so that's what you will get. If you somehow asked utl_file to open a file called myspool.log and you ended up with myspool.txt this would be a bug and you had best log this with Oracle Support. However, if you are somehow confusing the file extension (a naming convention foisted on us by Micro$oft) with the actual contents of a file (plain text ascii or binary) then the answer is, in 9i, utl_file only gives you the abilty to write text files. Writing binary is available in 10g.
There are workarounds to write binary files in 9i.
HTH
Chris
Message was edited by:
Chris Poole (because I can't spel)
Message was edited by:
Chris Poole (still can't spel)

How to create/delete files from filesystem using PL/SQL ? UTL_FILE?

Greetings,
I will start by explaining what i intend to do.
I have an application made in APEX. This application will have among other purposes the managment of pdf files which will reside in the filesystem.
I have questioned the person in charge to keep the pdf files in the database and not in the filesystem but without success.
So the pdf files reside in the filesystem and there is a record in a database table about them. A table keeps all info about the pdf, their location , size and name, creation date etc.
The APEX application will have a mecanism to allow the deletion of the pdf files if an administrator decides.
So it should be possible for an administrator to schedule the deletion of all pdf files whoe creation date is older than 2008 for example
So, how can i achieve that?
After some research i foudn about the UTL_FILE package which seems to have it takes to perform the task in issue.
My idea was to have a script in the operating system which runs nightly and reads a file containing all file names of the pdf to be erased.
The file which contains the names of the pdfs to be erased will be generated by the database a few minutes before.
If there are no pds files to be erased than the file containing the names will simply be empty
Are there any other viable solutions out there?
And as for opening/creating the file withn the pdf names, i use:
UTL_FILE.FOPEN (
location IN VARCHAR2,
filename IN VARCHAR2,
open_mode IN VARCHAR2,
max_linesize IN BINARY_INTEGER)
RETURN file_type;
And as for writing lines (a pdf name per line ), i use;
UTL_FILE.PUT_LINE (
file IN FILE_TYPE,
buffer IN VARCHAR2,
autoflush IN BOOLEAN DEFAULT FALSE);
is there a better solution?
thanks all.
-> My Homepage <-
Edited by: Igor Carrasco on Apr 14, 2009 3:11 PM
Edited by: Igor Carrasco on Apr 14, 2009 3:12 PM

Greetings,
I have read that link above, some questions still though.
I will provide some more information.
-First the database is in a windows server.
The windows server has a virtual drive mounted as z:\ <-- this points to a directory in virtual machine, i can manually access/create/delete files manually,i tested.
-Second utl_file_dir is defined as * , in t that enough to cover mounted drives? ( i can't change the init.ora and reboot the db right now :( gotta wait.. )
Do i explicitly have to define utfl_file_dir = z: ?
-Third haven't had the chance to test it on linux or any other operating system, assuming a virtual unit is mounted successfully and that the issues above are solved i should be able to operate on any mounted drive whatever the os, right?
Best regards

Permissions on files created using UTL_FILE package

The files created on unix using UTL_FILE package have permisisons rw- - - - - - -.(600). I want them to be created with permissions 640.(rw-r - - - - - ).i.e read access to the group also. The umask setting of the unix account of the oracle instance are 137. Is there any way to create the files with the required permissions.

So SQL*Plus is on your PC? I'm assuming you mean 'call' as in the windows cmd.exe command?
Sheesh this is like getting blood from a stone.
In that case you must be connecting via the listener and not internally.
In which case the umask comes from the umask that was set in the environment of the OS user who started the listener process.
Thats not necessarily the same as the owner of the oracle software (normally 'oracle').
If sysdamin joebloggs logs in and starts the listener manually in a shell, then the umask applied to all shadow processes started by the listener is whatever jobloggs umask was when he typed the command 'lsnrctl start'. The umask of the owner of the oracle software, or your umask, is utterly irrelevent.
So if you don't like the umask for files created by processes spawned by the listener, stop and restart the listener in a shell where you have explicitly set the umask to what you want it to be.
Really starting the listener should be scripted and the correct umask is put in the script just before the call to 'lsnrctl start'.

Does we need to Create a file Before using Open dataSet

Hi Guys,
            Can any body tell me whether we have to create a File before using Open dataset for writing files in Application Server.
           Can anybody write a program for EKPO table to store it in application Server using OpendatasSet ?Is there any specific care we have to take for Executing   pgm in Background?
I wrote the pgm for Foreground using GUI- Download.
PGM.
      select * into table it_ekpo from ekpo .
CALL FUNCTION 'GUI_DOWNLOAD'
EXPORTING
filename ='c:\EKPO.xls'
filetype = 'ASC'
WRITE_FIELD_SEPARATOR = 'X'
TABLES
data_tab = it_ekpo
*fieldnames = l_heading
EXCEPTIONS
file_write_error = 1
no_batch = 2
gui_refuse_filetransfer = 3 .
                   Plzzzzzzzzz Make the changes for existing Pgm.any help is awarded with points.Its Urgent.
Thanks,
Gopi.

Hi Guys,
             Thanks for ur immediate response.I am sending the code just confirm me whether is correct r not?
DATA : wa_EKPO_file   TYPE file_table-filename.
select * into table it_ekpo from ekpo .
OPEN DATASET wa_EKPO_file FOR OUTPUT         "Write to appl. server
                                      IN TEXT MODE
                                      ENCODING DEFAULT. "Open dataset Return code
IF sy-subrc = 0.
LOOP at IT_EKPO into WA_EKPO.
TRANSFER wa_ekpo to wa_ekpo_file.
ENDLOOP.
ENDIF.
                                         I want write the data into SPREAD Sheet .Can u guys tell me how to write into Spread Sheet using Open data Set.
Thanks,
Gopi.

How do I use UTL_FILE to insert a large number of fields to a file?

Hi
I am trying to use UTL_FILE for the first time in a Stored Procedure. I need to run a complex query to select 50 fields from various tables. I need these to be inserted into one line in the output file for all rows. Is this possible? My procedure so far is like the following
CREATE OR REPLACE PROCEDURE PROC_TEST IS
output_file UTL_FILE.FILE_TYPE;
BEGIN
FOR query in (SELECT FIELD1, FIELD2, ..........FIELD50)
FROM TABLE A, TABLE B
WHERE A.ID = B.ID
ETC
LOOP
UTL_FILE.PUT_LINE(output_file, <put all 50 fields for all records into file> );
END LOOP;
UTL_FILE.FCLOSE (output_file);
EXCEPTION
WHEN NO_DATA_FOUND THEN
NULL;
WHEN OTHERS THEN
UTL_FILE.FCLOSE_ALL;
RAISE;
END PROC_TEST;
Do I need to define 'query' (after the FOR) anywhere, also please advise with how I put all of the fields into the file.
Thanks
GB

Thanks Steve,
I have the UTL_FILE working fine now.
I have other queries to run and conditions to apply in the same procedure, and I need to schedule via Enterprise Manager, therefore using UTL_FILE in a procedure seemed the best option. I looked up Data-pump but this seems to be an 11g feature, and we are still on 10g therefore I will not be able to use it.
Thanks for your help.
GB

How to read a tab seperated data from a text file using utl_file

Hi,
How to read a tab seperated data from a text file using utl_file...
I know if we use UTL_FILE.get_line we can read the whole line...but i need to read the tab separated value separately.....
Thanks in advance...
Naveen

Naveen Nishad wrote:
How to read a tab seperated data from a text file using utl_file...
I know if we use UTL_FILE.get_line we can read the whole line...but i need to read the tab separated value separately.....If it's a text file then UTL_FILE will only allow you to read it a line at a time. It is then up to you to split that string up (search for split string on this forum for methods) into it's individual components.
If the text file contains a standard structure on each line, i.e. it is a fixed delimited structure, then you could use external tables to read the data instead.

HOW TO READ DATA FROM A FILE AND INSERT INTO A TABLE USING UTL_FILE

Hi..
I have a file.I want to read the data from file and load it into a table using utl_file.
how can I do it?
Any reply apreciated...

Hi,
This is not your requirment but u can try this :
CREATE OR REPLACE DIRECTORY text_file AS 'D:\TEXT_FILE\';
GRANT READ ON DIRECTORY text_file TO fah;
GRANT WRITE ON DIRECTORY text_file TO fah;
DROP TABLE load_a;
CREATE TABLE load_a
(a1 varchar2(20),
a2 varchar2(200))
ORGANIZATION EXTERNAL
(TYPE ORACLE_LOADER
DEFAULT DIRECTORY text_file
ACCESS PARAMETERS
(FIELDS TERMINATED BY ','
LOCATION ('data.txt')
select * from load_a;
CREATE TABLE A AS select * from load_a;
SELECT * FROM A
Regards
Faheem Latif

Issue while accessing a file using UTL_FILE

Hi,
My requirement is to check for the existence of a file in the file system(apps tier) and then launch a conc program to load the data.
There is a separate DB tier.
I am trying to use UTL_FILE.fopen to indirectly see if the file exists
and then proceed and I am trying to use UTL_FILE.fopen for that.
The file is placed in the apps tier. I have created a directory in the database with the path and using that in the UTL_FILE.FOPEN call.
However I am getting the following error. ORA-29283 - An attempt was made to read from a file or directory that does not exist, or file or directory access was denied by the operating system.
What could be the issue here? Is there anything that I need to do after creating the directory?.
Thanks,
Balaji

/Pl post details of OS, database and EBS versions.
Are the Apps and DB servers physically separate ? If so, the directory on the Apps server where the file resides will have to be visible/updateable to the DB server in order for UTL_FILE to be able to read/write to it.
HTH
Srini/
The apps and db servers are separate. EBS version is 11.5.10.2 and the OS is unix.
How do I make the directory in apps server visible/updateable to the DB server?. I have created a directory in the database with the location of the file in apps tier.
Thanks for your quick reply
Regards
Balaji

Concatenating file using utl_file pkg.

Hi ,
I have files generated every 10 mins. I have to concatenate all these files for every 5 hours and distribute that file to client locations.
can u plz. tell me is it possible to concatenate small but huge number of files
using utl_file package.
Help needed !!
Thanks in advance

Hi Guido
Thanks for the reply ..
But can i call this shell script from oracle procedure ...
As i have this files in DB server and i don't have my java there ...
I have to invoke this shell script from oracle ...

Flush file before writing using UTL_FILE

Similar Messages

Maybe you are looking for