Re: Characterset Question

When you use datapump, it's not an issue with NLS_LANG. Datapump is taking care of that.

Your problem with the migration of US7ASCII databases to AL32UTF8 is, as you say, that some characters that are stored in 1 byte in US7ASCII, they take up 2 bytes when stored in AL32UTF8. So when a VARCHAR2 column has a length of say 20 *bytes*, it can contain anywhere from 5 to 20 *characters *of UTF8, depending on the byte-length of each character (between 1 and 4.)
In UTF8 database, instead of defining character column length in bytes, they should be defined in characters.

The fix for this is to change your columns from for example VARCHAR2(20) to VARCHAR2(20 *CHAR*).
When the column is defined as VARCHAR2(20 *CHAR*), it can contain 20 *characters*, no matter if they use 1 or 2 or 3 or 4 bytes of storage.

Basically for every VARCHAR2 column in the database, do an ALTER TABLE MODIFY (<colname> VARCHAR2(<length> *CHAR*)); (The DMU can also help to do this - change all columns to use CHAR semantics instead of BYTE semantics.)

I think you'd need to do this change to CHAR semantics in the source databases, so that then impdp creates the tables, they'll be created correctly.
I don't think it'll work to try and force the impdp to create the tables with CHAR semantics by setting target database parameter NLS_LENGTH_SEMANTICS to CHAR before doing the impdp, but maybe give it a try on a small testcase?

Cheerio
/Kim

On Mon, Mar 25, 2024 at 3:40 PM Scott Canaan <dmarc-noreply_at_freelists.org> wrote:

> Catching up on these messages.
>
>
>
> The characterset of the source (PeopleSoft) database is AL32UTF8. The
> characterset of the ultimate destination (Data Warehouse) database is
> AL32UTF8. In between them is Informatica. That is running in Windows
> using a SQL Server database. It used to be Oracle, but my previous boss
> decided that he wanted it using SQL Server, so he had it converted. I
> don’t have access to the Informatica app server. I suspect that’s where
> the issue is.
>
>
>
> Another related issue is that we are in the process of migrating from Red
> Hat 7 to Red Hat 8. Most of the databases in Red Hat 7 are US7ASCII, but
> the new databases in Red Hat 8 are AL32UTF8, as it is the default now. If
> there is a “special” character in the US7ASCII database, when it is
> datapumped into the Red Hat 8 database, it takes 2 bytes instead of one,
> which sometimes causes the import to error with an “ORA-12899: value too
> large for column” error. I made sure the environment running the impdp
> command (over the network) has the NLS_LANG set to
> AMERICAN_AMERICA.AL32UTF8. I’m not sure how to fix that without going
> backward to US7ASCII, which I don’t want to do.
>
>
>
> *Scott Canaan ‘88*
>
> *Sr Database Administrator *Information & Technology Services
> Finance & Administration
>
>
> *Rochester Institute of Technology *o: (585) 475-7886 | f: (585) 475-7520
>
> *srcdco_at_rit.edu <srcdco_at_rit.edu>* | c: (585) 339-8659
>
> *CONFIDENTIALITY NOTE*: The information transmitted, including
> attachments, is intended only for the person(s) or entity to which it is
> addressed and may contain confidential and/or privileged material. Any
> review, retransmission, dissemination or other use of, or taking of any
> action in reliance upon this information by persons or entities other than
> the intended recipient is prohibited. If you received this in error, please
> contact the sender and destroy any copies of this information.
>
>
>
> *From:* Mark W. Farnham <mwf_at_rsiz.com>
> *Sent:* Friday, March 22, 2024 6:56 PM
> *To:* dimensional.dba_at_comcast.net; joncrisler_at_gmail.com;
> Clay.Jackson_at_quest.com
> *Cc:* Scott Canaan <srcdco_at_rit.edu>; oracle-l_at_freelists.org
> *Subject:* RE: Characterset Question
>
>
>
> If memory serves, the utility was csscan, and I **thought** it was
> upgraded at some point to flag bytes not (supposed to be) in the source
> database characterset.
>
>
>
> That’s probably moot for this case, since Scott identified that inserts
> from informatica into the new server are the apparent problem, and as Clay
> pointed out, checking the NLS_LANG (I think) on the client running
> informatica should be checked.
>
>
>
> If the client and server nls language values versions are identical, I *
> *think** no translation is done, which if Scott is correct they are both
> AL32UTF8, then the translation is the no-op, which should both work
> correctly and be infinitesimally faster.
>
>
>
> Scott, at your convenience please let us know how you make out.
>
>
>
> *From:* dimensional.dba_at_comcast.net [mailto:dimensional.dba_at_comcast.net
> <dimensional.dba_at_comcast.net>]
> *Sent:* Friday, March 22, 2024 6:03 PM
> *To:* joncrisler_at_gmail.com; Clay.Jackson_at_quest.com
> *Cc:* srcdco_at_rit.edu; 'Mark W. Farnham'; oracle-l_at_freelists.org
> *Subject:* RE: Characterset Question
>
>
>
> The tool works great all charactersets except for databases that are set
> to us7ascii and have 8bit bytes too.
>
>
>
> *From:* oracle-l-bounce_at_freelists.org <oracle-l-bounce_at_freelists.org> *On
> Behalf Of *Jon Crisler
> *Sent:* Friday, March 22, 2024 2:22 PM
> *To:* Clay.Jackson_at_quest.com
> *Cc:* srcdco_at_rit.edu; Mark W. Farnham <mwf_at_rsiz.com>;
> oracle-l_at_freelists.org
> *Subject:* Re: Characterset Question
>
>
>
> There is a utility that is part of the Oracle characterset conversion
> program, that can scan the existing DB and the proposed characterset. It
> will report back on any problem tables / columns and if the result is
> lossless, losey and/or info about manual cleanup. The exact name escapes
> me, but check out Oracle DocIDs relating to converting from one
> characterset to another and it should be mentioned. I will try to find it
> as well.
>
>
>
> On Fri, Mar 22, 2024 at 4:38 PM Clay Jackson <dmarc-noreply_at_freelists.org>
> wrote:
>
> Check the Oracle client NLS settings that the Informatica server is
> using. 😊
>
>
>
>
>
> *Clay Jackson*
>
> Database Solutions Sales Engineer
>
> <https://www.quest.com/solutions/database-performance-monitoring/>
>
> clay.jackson_at_quest.com
>
> *office* 949-754-1203 *mobile* 425-802-9603
>
>
>
> *From:* oracle-l-bounce_at_freelists.org <oracle-l-bounce_at_freelists.org> *On
> Behalf Of *Scott Canaan
> *Sent:* Friday, March 22, 2024 11:54 AM
> *To:* Mark W. Farnham <mwf_at_rsiz.com>; oracle-l_at_freelists.org
> *Subject:* RE: Characterset Question
>
>
>
> *CAUTION:* This email originated from outside of the organization. Do not
> follow guidance, click links, or open attachments unless you recognize the
> sender and know the content is safe.
>
>
>
> Interesting that the source is also AL32UTF8, but the data is transferred
> via Informatica. My guess is that’s where the issue is.
>
>
>
> *Scott Canaan ‘88*
>
> *Sr Database Administrator *Information & Technology Services
> Finance & Administration
>
>
> *Rochester Institute of Technology *o: (585) 475-7886 | f: (585) 475-7520
>
> *srcdco_at_rit.edu <srcdco_at_rit.edu>* | c: (585) 339-8659
>
> *CONFIDENTIALITY NOTE*: The information transmitted, including
> attachments, is intended only for the person(s) or entity to which it is
> addressed and may contain confidential and/or privileged material. Any
> review, retransmission, dissemination or other use of, or taking of any
> action in reliance upon this information by persons or entities other than
> the intended recipient is prohibited. If you received this in error, please
> contact the sender and destroy any copies of this information.
>
>
>
> *From:* Mark W. Farnham <mwf_at_rsiz.com>
> *Sent:* Friday, March 22, 2024 2:50 PM
> *To:* Scott Canaan <srcdco_at_rit.edu>; oracle-l_at_freelists.org
> *Subject:* RE: Characterset Question
>
>
>
> That should be a lossless conversion with the possible exception of
> clob/blob.
>
>
>
> However, figuring out the appropriate NLS_PARAMETER values for the client
> and the server have to be set right.
>
>
>
> How are you bringing data from Peoplesoft over to your new database?
>
>
>
> IF it is application to database, setting the client and server parameters
> should take care of it.
>
>
>
> IF it is database (old) to database (new), then there is possibly some
> data in the (old) database that is not actually US7ASCII and some cleanup
> may be required.
>
>
>
>
> https://docs.oracle.com/en/database/oracle/oracle-database/19/nlspg/supporting-multilingual-databases-with-unicode.html
>
>
>
> should give you a leg up, and there are more similar documents on Oracle.
>
>
>
> I’m a bit rusty on this, my friend, but I think that is about correct.
>
>
>
> mwf
>
>
>
>
>
>
>
> *From:* oracle-l-bounce_at_freelists.org [
> mailto:oracle-l-bounce_at_freelists.org <oracle-l-bounce_at_freelists.org>] *On
> Behalf Of *Scott Canaan ("srcdco")
> *Sent:* Friday, March 22, 2024 2:23 PM
> *To:* oracle-l_at_freelists.org
> *Subject:* Characterset Question
>
>
>
> We are in the process of migrating all of our Oracle databases from Red
> Hat 7 to Red Hat 8. In creating the new databases in Red Hat 8, we’ve been
> going with the default characterset of AL32UTF8. Many of the old databases
> are US7ASCII. We are having trouble with applications trying to insert
> into text fields in the new database. In one case, bringing data from
> Peoplesoft that has a degree symbol fails because that comes over as 2
> bytes instead of one. I thought that AL32UTF8 should handle those
> characters. What needs to be done to accommodate this?
>
>
>
> *Scott Canaan ‘88*
>
> *Sr Database Administrator *Information & Technology Services
> Finance & Administration
>
>
> *Rochester Institute of Technology *o: (585) 475-7886 | f: (585) 475-7520
>
> *srcdco_at_rit.edu <srcdco_at_rit.edu>* | c: (585) 339-8659
>
> *CONFIDENTIALITY NOTE*: The information transmitted, including
> attachments, is intended only for the person(s) or entity to which it is
> addressed and may contain confidential and/or privileged material. Any
> review, retransmission, dissemination or other use of, or taking of any
> action in reliance upon this information by persons or entities other than
> the intended recipient is prohibited. If you received this in error, please
> contact the sender and destroy any copies of this information.
>
>
>
>

--
http://www.freelists.org/webpage/oracle-l







(image/jpeg attachment: image001.jpg)