Accented (Hungarian) characters - puzzle.

Question

Hi all,
I have a .csv file with English words and their Hungarian equivalent - the .csv file is
fine and appears with the correct accents in both LibreOffice Calc and with gedit
and the command line (i.e. $> more my_file.csv) in bash.
==========================
consumption,fogyasztás
consumption,tüdővész
precious,értékes
precious,drága
===========================
I then create an external table
=================================
CREATE TABLE Hungarian_dict
Angol VARCHAR2(50),
Magyar VARCHAR2(50)
ORGANIZATION EXTERNAL
type oracle_loader
default directory MY_FILES
access parameters
records delimited by newline
fields terminated by ','
missing field values are null
location ('SZOTAR.csv')
reject limit unlimited;
============================
The problem is, for example, the word "consumption"
Select * from hungarian_dict where angol = 'consumption';
appears thus (exported from SQL Developer (also gibberish in SQL*Plus).
==========================
pol@pol-laptop:~/Downloads/personal/hungary$ more export.tsv
"ANGOL"     "MAGYAR"
"consumption"     "fogyasztás"
"consumption"     "tüdÅ‘vész"
pol@pol-laptop:~/Downloads/personal/hungary$
===========================
My question is how do I get Oracle to display the correct settings? Do I have to
reload my table with different NLS_LANG settings? NLS_LANGUAGE in the
database?
===========
SQL> show parameter NLS_LANGUAGE;
NAME                    TYPE     VALUE
nls_language               string     AMERICAN
SQL>
===============
at the bash prompt, if I type "locale" the result is
=====================================
en_IE.utf8
=====================================
for everything.
Should I create my table with a different character set and collation for the
Hungarian language field?
TIA,
Paul...

Answer 1

>
Firstly, thanks to both Srini and Orafad - just a followup (puzzling).
NLS_CHARACTERSET     WE8MSWIN1252
Without CHARACTERSET clause, datafile encoding need to match database character set.
But I guess data you have is in UTF-8.OK, I put in CHARACTERSET UTF8 in my access parameters for my table.
I also specified NVARCHAR for my input fields.
This appears to give the correct characters - i.e. I ran these queries
SELECT * from angol_magyar where angol = 'handkerchief';
SELECT * from angol_magyar where angol = 'glove';I get these results
"ANGOL"                  "MAGYAR"
"handkerchief"           "zsebkendő"      <- note o-double-acute - no problems with either SQL Developr or exporting to text - more in bash OK.
"glove"                       "kesztyű"         <- note u-double-acute -                           ditto.But, when I run these queries
SELECT * from angol_magyar where magyar LIKE '%ő%';
SELECT * from angol_magyar where magyar LIKE '%ű%';i.e. with the o-double-acute or the u-double-acute surrounded by % signs - which I thought would
just bring back all the words with the double-acute accented characters, these queries in fact bring back
any magyar word with o - i.e. "o" on its own and o-acute and o-umlaut and o-double-acute.
This is puzzling to say the least. Any ideas?
To verify data stored, you could use:
select column, dump(column, 1016) from table where suitable_condition_to_pick_row ... ;For the query
SELECT magyar, dump(magyar, 1016) from angol_magyar where (angol = 'handkerchief' or angol = 'glove');
The result is (gibberish to me):
"kesztyű"     "Typ=1 Len=14 CharacterSet=AL16UTF16: 0,6b,0,65,0,73,0,7a,0,74,0,79,1,71"
"zsebkendő"     "Typ=1 Len=18 CharacterSet=AL16UTF16: 0,7a,0,73,0,65,0,62,0,6b,0,65,0,6e,0,64,1,51"Also, how do I have different collations for the English and Hungarian fields in my table?
Paul...