sort characters in a varchar2

Similar Questions

• VARCHAR2 is impaired

  
  Hello everyone,

  I am a developer concerned on the following problem, I have in my recent copy of data.

  I create a table using a CREATE TABLE AS SELECT OPTION in a 11.2 g database located on RedHat Linux operating system.

  the table is on the 10.2 on Solaris database, operating system, the problem is that the varchar2 data type is changed in the target database.

  for example

  data:

  Field1 varchar2 (2)

  to the data

  Field1 varchar2 (6)

  all varchar2 fields are translated as I have explained before. (expanded 3 times of the original field size is)

  data types: number, float, date of: no problem

  I use a dblink in order to access the data and my Oracle client is 10.2.0 version.

  Colud you help me to solve this problem,

  I appreciate any idea that you should have.

  Kind regards.

  By default, if you declare a column VARCHAR2 (2), you are allocating 2 bytes of storage.  In your data base ISO-8859-1, which is enough to store 2 characters, because each character requires 1 byte of storage.  In your database to UTF-8, however, which may not be enough store a single character since UTF-8 is a variable-width character set and each character it will take between 1 and 3 bytes of storage.

  If you declare a column VARCHAR2 (2 CHAR), on the other hand, you use the semantics of length in characters rather than the semantics of length in bytes.  Which allocates space for 2 characters regardless of whether it requires 2 or 6 bytes of storage under the covers (subject to the same limit of 4000 bytes the maximum size of a VARCHAR2).  As a general rule, if you develop a UTF-8 database, you want to be using character length semantics.  Which may involve specifying way explicit semantics of length in characters on each VARCHAR2 column.  It may involve making an ALTER SESSION set NLS_LENGTH_SEMANTICS to CHAR before executing the CREATE TABLE statement.  Whatever it is, you'll probably want to do first your CREATE TABLE and then run an INSERT.

  If you do not use the semantics of length in characters, you will need to increase the size of your columns in order to ensure that you have a sufficient space in the database to UTF-8 for all your data.

  Justin

• space of VARCHAR2

  Hi all
  I need to store more than 4000 space characters in a column varchar2 oracle database 10g.
  
  can any one guide me how to store more than 4000 characters in a varchar2 column...
  
  any help is deeply appreciated.

  user12284380 wrote:
  Hi all
  I need to store more than 4000 space characters in a column varchar2 oracle database 10g.

  can any one guide me how to store more than 4000 characters in a varchar2 column...

  any help is deeply appreciated.

  Yes, use a CLOB data type.

• difference between the view and the procedure

  Hi Master,

  Today I received a request from my senior technical leader... !!

  Q > what is the difference between the view and the procedure?

  I got response like this...

  A view represents the logic of one or more tables/views. It won't take no place in db... It is a virtual table, query will only store in db... useful

  for safety... If the main table has obtained changes... automatically discovers also change... etc.

  Proecudure is a db object will be performs a perticular action... It may or may not return values. ?

  But somehow, he isn't happy... He expects more... .very close view of diff. b & w and procedure.

  Please explain it to me...!

  Concerning

  AR

  874273 wrote:

  Hi Blu,

  Thanks for your replies...! But we think if notice and procedure will be the same no matter what point of view? Both are the same? If Design view... We can create a stored procedure?

  How a view and procedure may do the same thing?  They are two different things with different objectives.  You use the correct one depending on what you want to achieve; they are not interchangeable.

  And another question...!

  (Q) how you will store Japan, characters of China in a database table? will there be a new feature in Oracle 11 g?

  (A) I use dbms_lob package? Japan/China characters are multi byte characters... so we can use the nclob data type?

  is this correct... ?

  No, it not there no novelty in Oracle 11 g for the storage of Japanese or Chinese characters, because there is already a feature for their storage in previous versions as well.  It is determined by the parameters characterset and nationality that you use when you create the database.

  Why would you choose to store them in a LOB using DBMS_LOB?  Yes, they have multibyte characters, but if you have configured your database for the correct character sets, for example using UTF - 8, then you can store the characters in VARCHAR2, like any other language.  The only difference is that the storage of 4000 bytes limit still applies, so you maybe isn't able to store 4000 multibyte characters in a varchar2 column 4000 single-byte, it can be as little as 1000 characters if they are all characters in 4 bytes.  (Note: 12 c limit in SQL can be increased up to 32,767 characters, to harmonize it with variable varchar2 PL)

• Forming a sql statement

  Hey,.
  
  I have a small doubt not a large...
  
  Forming the dynamic sql code.
  
  -I have to train a SQL like this
  
  Select * from t1
  Union
  Select * from t2
  Union
  Select * T3...
  
  I could train sql but an additional union is added at the end.
  I am unable to use rtrim my intervention is more than 4,000 characters.
  
  
  declare
  V_Sql Varchar2 (32767).
  V_TBL VARCHAR2 (32767).
  V_Sql1 Varchar2 (32767).
  v_sql2 varchar2 (32767).
  
  cursor c1 is select table_name from user_tables where table_name like '% t';
  
  Start
  for the l1 c1 loop
  v_tbl: = l1.table_name;
  v_sql: = ' Select * from '. v_tbl | "the union";
  v_sql1: = v_sql1 | v_sql;
  End loop;
  Insert Into X values (V_Sql1);
  commit;
  end;
  
  
  
  
  
  Thank you

  user13309379 wrote:
  I am unable to use rtrim my intervention is more than 4,000 characters.

  You do not want to use RTRIM anyway; It would remove any 'U,' n, ' I has or ' O that happened at the end of the name.
  You want to use SUBSTR, and (at least in Oracle 10 and more) it works very well on VARCHAR2s more than 4000 characters:

  set     serveroutput     on
  
  declare
       V_Sql1     Varchar2 (32767) := RPAD ('X', 20000, 'Y');
  begin
       dbms_output.put_line (LENGTH (v_sql1) || ' = LENGTH before');
  
       v_sql1 := SUBSTR ( v_sql1
                  , 1
                  , LENGTH (v_sql1) - 7
                  );
  
       dbms_output.put_line (LENGTH (v_sql1) || ' = LENGTH after');
  end;
  

  produces this output:

  20000 = LENGTH before
  19993 = LENGTH after
  

• Sort problem: ORDER BY the numbers all first instead of characters first

  Hello
  
  I need help with the problem of the following sort:
  
  create table test01 (sorting_column varchar2 (30));
  
  insert into values test01 ('Test123');
  insert into values test01 ("Test987");
  insert into values test01 ("123Test");
  insert into values test01 ("987Test");
  insert into test01 values (string 'test');
  
  Select sorting_column from test01 arrested by sorting_column;
  
  SORTING_COLUMN
  ------------------------------
  Test string
  Test123
  Test987
  123Test
  987Test
  
  This is not the order that I need. I really wish it were (numbers before characters):
  
  123Test
  987Test
  Test123
  Test987
  Test string
  
  Can anyone help?
  
  Thanks in advance,
  Marc

  SELECT sorting_column
  FROM test01
  ORDER BY nlssort(sorting_column, 'NLS_SORT=BINARY')
  ;
  

  Can be set at the level of the session as well:

  ALTER SESSION SET NLS_SORT = BINARY;
  

  Published by: odie_63 on June 20, 2011 17:45

• Problem sorting by column > 500 characters in length when NLS_SORT is GENERIC_M

  Hi all
  
  I ran into this somewhat misleading error message, a DDL, an institution which, when I run a DML statement. What I try to do is, I set everything first NLS_LANGUAGE = NLS_SORT "AMERICAN" = "GENERIC_M" values for the session and then retrieve the data from a table, you place your order by a column called description_value (varchar2 (4000)).
  
  Like this
  
  ALTER SESSION SET NLS_LANGUAGE = NLS_SORT "AMERICAN" = "GENERIC_M."
  
  Select * from tableA by description_value
  
  The light of the foregoing, I execute via Java, but I am able to run roughly in the same error message in TOAD/PL-SQL developer. The error message is as below,
  
  ORA-00910: specified length too long for its data type
  
  _ Some points:
  
  1. I get this problem only when there are more than 500 record characters with the length of the column description_value. Anything less than 500 characters, I don't get the error message.
  2. when I put the NLS_SORT 'BINARY' or 'FRENCH', I do not run on this issue. The query runs correctly.
  3. I never had this problem in 10 g. This question arose in 11.
  
  I did a bit of google but din't find any specific reference to anyone else having this problem. Please share your thoughts and comments on this and please let me know if you need more information about it.
  
  Published by: user13405584 on December 2, 2010 04:20

  Hello

  I was too quick to research. It is indeed wrong. You should read ""Bug 9723547: THE LIMITATION OF the FUNCTION NLSSORT '. " This bug describes the behavior of the NLSSORT and there are bugs in mentioned to correct a wrong implementation of a bug sooner. Because you do not give your exact version of Oracle, take a look and see a patch that you need to correct the false correction of the bug.

  Besides in this bug, it is also clear, documentation was not correct. This is now fixed in the Guide of globalization: kinds of Linguistics [url http://download.oracle.com/docs/cd/E11882_01/server.112/e10729/ch5lingsort.htm#insertedID6], under classification keys. There is the claim that a sort on a kind of multi-langual can be on about 500 characters.

  Read the bug and solutions (patches), and I hope that this becomes clearer.

  Herald tiomela
  http://htendam.WordPress.com

• Region - Sorting Varchar2 as report number

  I have reports that contain a column with several decimal places in it. Due to several decimal places I use Varchar2 data type.
  
  The problem is that users want to sort on this column and its sort numbers as characters. I don't think I can use the "order by lpad" solution that I had in a normal query because I got an error when I tried
  
  Example of how APEX sorts - HIGH to LOW
  3.0.1
  14.0.0
  13.9.9
  
  How do I sort - HIGH to LOW
  14.0.0
  13.9.9
  3.0.1
  
  
  Has anyone experience this problem?

  I'm not sure what is the problem, but try this, add another ip_address column (with alias IP_ADDRESS2) and make a hidden column and do the IP_ADDRESS_SORT to use #IP_ADDRESS2 # instead of #IP_ADDRESS #...

  Sam

• Sort numerically a varchar2 container field '-' and tanks

  I have a Field1 varchar2 (20) that contains values such as
  2-14', 3-9', '1-1-2 T'
  
  They should be sorted in the order
  
  ' 1 1-2 T '
  2-14'
  3-9'
  
  I tried to use lpad (field1, 22, '0') but he left buffers at a given fixed length for the whole area and it does not work.
  
  There is no fixed grammar for this field and it is possible that in the future some unexpected tank can appear in the field. For example, "[1-2]" "
  
  But the query needs to work even if the sort is not correct.
  
  

  select field1,field2,field3...from table tab1, tab2...where ...
            order by field1;

  Can someone please help?
  
  Thank you.

  Hello

  Sorry, I don't understand the sorting rules.
  It looks like
  (a) the string is divided into a number of secondary channels by hyphens.
  (b) sorting through secondary channels in the order, i.e. the substring n matter only if there is equality of strings 1 through n - 1.
  (c) strings composed entirely of numbers preceded by secondary channels who have other characters
  (d) secondary digital channels are classified as numbers, that is, '1' and '01' are the same, but comes from '1' before ' 03 "
  (e) other substrings are classified as strings

  Under these rules, the two identical channels ' 1 - 1 - R 'would come after the strings that start with 1-01 - 1-'.
  Other than that, the following query is what you asked:

  WITH  cntr     AS
  (
       SELECT     LEVEL     AS n
       FROM     dual
       CONNECT BY  LEVEL <= 1 + ( SELECT  MAX ( LENGTH (txt)
                                                - LENGTH ( REPLACE ( txt
                                                      , '-'
                                            )
                                     )
                                )
                         FROM    table_x
                       )
  )
  ,       got_sub_txt  AS
  (
       SELECT  id
       ,     txt
       ,     n
       ,     REGEXP_SUBSTR ( txt
                       , '[^-]+'
                       , 1
                       , n
                       )          AS sub_txt
       FROM    table_x
       JOIN     cntr     ON  n <= 1 + LENGTH (txt) - LENGTH (REPLACE (txt, '-'))
  )
  ,     got_sort_key     AS
  (
       SELECT     got_sub_txt.*
       ,     SYS_CONNECT_BY_PATH ( CASE
                             WHEN  REGEXP_LIKE ( sub_txt
                                                 , '[^0-9]'
                                           )
                             THEN  '2' || RPAD (sub_txt, 10)     -- Magic Number
                             ELSE  '1' || LPAD (sub_txt, 10, '0')     -- Magic Number
                            END
                          , '-'
                          )     AS sort_key
       FROM     got_sub_txt
       WHERE   CONNECT_BY_ISLEAF = 1
       START WITH  n     = 1
       CONNECT BY  n     = PRIOR n + 1
       AND         id     = PRIOR id
  )
  SELECT    id
  ,       txt
  FROM       got_sort_key
  ORDER BY  sort_key;
  

  This requires that each substring will be (at most) 10 characters long. That's what the 'magic number' 10 in got_sort_key of the subquery. You can change this to whole m that you like.
  The only limit is the number of subkeys is that they must all fit in a VARCHAR2 4000 characters, so the limit is 4000 / (m + 2) (e.g., 333 when m = 10).

  For this solution, you need a unique identifier (id in the example above) for each line. You don't need to display this identifier in the SELECT clause.
  If txt itself is unique, it can be used as the identifier.

• SQL for sort a column containing digital/characters/alphanumeric characters.

  I have a table with the column can contain numbers, characters, and a combination of numbers and characters. Here is an example of the table. In this regard, I would like to sort the table in the following way.
  
  (a) sort first numbers in numerical order and then would sort the others in the order of the characters.
  
  Example: -.
  
  RUN SQL
  SQL > select * from temp;
  
  ID
  ---------------
  A
  22
  46
  CC
  BBB
  46jkb
  1
  10
  100
  one
  
  AFTER CORRECTION
  SQL > < MAGIC SQL statement >
  
  ID
  ---------------
  1
  10
  22
  46
  46jkb
  100
  one
  A
  BBB
  CC
  
  I tried asciistr, to_char and to_number and many other combinations. However, no luck :-(. Any help will be much appreciated.
  
  SQL statement - to create and insert data in the table-
  create table temp (id varchar2 (15));
  insert into temporary values ('A');
  insert into temporary values ('22');
  insert into temporary values ('46');
  insert into temporary values ("CC");
  insert into temporary values ("BBB");
  insert into temporary values ('46jkb');
  insert into temporary values ('1');
  insert into temporary values ('10');
  insert into temporary values ('100');
  insert into temporary values ('a');
  commit;

  with t as (
             select 'A' str from dual union all
             select '22' str from dual union all
             select '46' str from dual union all
             select 'CC' str from dual union all
             select 'BBB' str from dual union all
             select '46jkb' str from dual union all
             select '1' str from dual union all
             select '10' str from dual union all
             select '100' str from dual union all
             select 'a' str from dual
            )
  select  str
    from  t
    order by lpad(regexp_substr(str,'^\d+'),max(length(regexp_substr(str,'^\d+'))) over(),'0') || upper(regexp_replace(str,'^\d+'))
  /
  
  STR
  -----
  1
  10
  22
  46
  46jkb
  100
  A
  a
  BBB
  CC
  
  10 rows selected.
  
  SQL> 
  

  SY.

• Internal sorting/replacing characters

  I have a column of scientific names that looks like this:

  Canis lupus

  Savin ARGeo (haguel)

  Plumeria algers

  I want to remove all the brackets and their content, so the finished product looks like this:

  Canis lupus

  ARGeo savin

  Plumeria algers

  However, I would first sort the column so that all the scientific names that includes parentheses are grouped. Then I can create a new column and a checkbox to indicate the values that previously contained in parentheses.

  So I was just wondering if there is a way to 1) sort columns according to the different values of the special characters contain and 2) remove all parentheses and all that they contain. And I would add that the brackets attach all "haguel." There are a lot of different words between brackets.

  Thank you.

  You can do without sorting by following this example:

  

  In the column A is the original name

  B1 = SIERREUR (FIND ("(", A1), 0 "))

  It's shorthand dethrone select cell B1 and type (or copy and paste it here) the formula:

  SIERREUR = (FIND ("(", A1), 0 "))

  C1 = IFERROR (Find (",") ", A1, start-pos), 0)

  D1 = if (AND (C1 > 0 > 0, B1), (A1, B1−1) LEFT & RIGHT (A1, LEN (A1) −C1), A1)

  Select cells B1 to D1, copy

  Select cells B1 at the end of column D, dough

  D column now contains the "corrected" names  Select copy of column d.

  now select the destination and use the menu item "Edition > paste formula results.

• Sort by or find the word long in a field of varchar2

  Oracle Database 10g Enterprise Edition Release 10.2.0.4.0 - 64bi
  PL/SQL Release 10.2.0.4.0 - Production
  CORE     10.2.0.4.0     Production
  TNS for HPUX: Version 10.2.0.4.0 - Production
  NLSRTL Version 10.2.0.4.0 - Production

  drop table event;
  create table event  (event_title_nm varchar2(1000));
  
  SET DEFINE OFF;
  Insert into EVENT
     (EVENT_TITLE_NM)
   Values
     ('CREATING A NEW CLASS WORD FOR INDICATOR, - _IN');
  Insert into EVENT
     (EVENT_TITLE_NM)
   Values
     ('Unknown Title');
  Insert into EVENT
     (EVENT_TITLE_NM)
   Values
     ('MASAL REASON TEXT');
  Insert into EVENT
     (EVENT_TITLE_NM)
   Values
     ('Note/Attachment Modifications and Amendments');
  Insert into EVENT
     (EVENT_TITLE_NM)
   Values
     ('Installation (replaces change request  66)');
  Insert into EVENT
     (EVENT_TITLE_NM)
   Values
     ('Unknown Title');
  Insert into EVENT
     (EVENT_TITLE_NM)
   Values
     ('Installation (replaces change request  66)');
  Insert into EVENT
     (EVENT_TITLE_NM)
   Values
     ('List Item Identifier Change');
  Insert into EVENT
     (EVENT_TITLE_NM)
   Values
     ('USER_ID in WORKFLOW_TASK_ASSIGNMENT');
  COMMIT;

  I would like to choose or order by the length of the longer term in the field
  Word defined as characters continued.
  
  so my production for the first two returned

  event_title_nm
  USER_ID in WORKFLOW_TASK_ASSIGNMENT
  Note/Attachment Modifications and Amendments

  because WORKFLOW_TASK_ASSIGNMENT is the largest word
  and Note/attachment is the second large.
  
  the reason why I was trying to find thiis is that I display this value in a web page and I need to understand about the longest
  continuous word because the css will not end a text in the middle of a Word. so, it ends up pushing my div set below where I want it to go.
  
  Published by: pollywog on October 25, 2011 08:43

  You can order by a scalar subquery returns the max length:

  SQL> select *
    2  from event e
    3  order by (
    4             select max(length(regexp_substr(e.event_title_nm, '\S+', 1, level)))
    5             from dual
    6             connect by level <= length(regexp_replace(e.event_title_nm, '\S+'))+1
    7           ) desc
    8  ;
  
  EVENT_TITLE_NM
  --------------------------------------------------------------------------------
  USER_ID in WORKFLOW_TASK_ASSIGNMENT
  Note/Attachment Modifications and Amendments
  Installation (replaces change request  66)
  Installation (replaces change request  66)
  CREATING A NEW CLASS WORD FOR INDICATOR, - _IN
  List Item Identifier Change
  Unknown Title
  Unknown Title
  MASAL REASON TEXT
  
  9 rows selected
   
  

• Problem copying 4000 characters of CLOB in VARCHAR2 (4000 TANK)

  I'm migrating data from a table where it is stored in a CLOB to another table where I said a column of VARCHAR2(4000 CHAR).
  
  I stated a cursor in the form:
  
  (a) select columna, columnb, substr(clob_column,1,4000) of...
  
  (b) select columna, columnb, dbms_lob.substr(clob_column,4000,1) of...
  
  I then loop and EXTRACT the value of substr in a variable that is declared as VARCHAR2 (4000 TANK)
  
  With the cursor version 'a', using the regular subtr function, I get: ORA-01461: can bind to a LONG value only for insert into a LONG column
  
  With the slider 'b' version, by using the dbms_lob.substr function, I get: ORA-06502: PL/SQL: digital or value error: character string buffer too small
  
  I use Oracle 10.2.0.4.0
  
  I don't know yet what to watch next?

  Machinery says:

  However, the database uses NLS_CHARACTERSET UTF8 and NLS_NCHAR_CHARACTERSET = AL16UTF16, which are multibyte so I was expecting the declared with semantic CHAR VARCHAR2 to be multibyte.

  And it is. But in Oracle SQL VARCHAR2, as I already mentioned, must not exceed 4000 bytes. So even if you have declared it as 4000 characters, you'll get an error when the total length in bytes is greater than 4000. And since, according to the character set, a character can be 1 to 4 bytes, Oracle takes the best cases and allows you to declare it as max 4000 characters. How many characters will fit in eseccion is another question.

  SY.

• Sort of a system of numbering both Alpha and numeric characters

  I have a database of about 370 products we manufacture. The numbering system is 3 digit design followed by the character alpha code 1-6 letters. For example:

  014A6LP (for our company, that means that number 14, while the A6 refers to a particular size and the LP of the design is the type of product).

  014FAP (same design, different product type)

  014A2LPBX (again, same design, different product type)

  I tried to sort my table such as numbers and text. I tried to add custom formats that include leading zeros. I even tried to let go the zeros (even if I really need them). I don't like to work.

  The result I want is this:

  001A6LP

  002A6LP

  002FAP

  003A2FPBX

  003A6LP

  003NBF

  etc, etc.

  what I get instead is this: as you can see it my list starts with 012 instead of 001. As it continues, 100 comes before 010, etc.

  

  I look forward to your ideas and help.

  Thank you

  Josh

  Hi Josh,.

  You can paste your table in a post (as opposed to a screenshot)?  In this way we can try the actual values you have.

  SG

• I have the blob column in a table. I want to break up the value of the column in characters varchar2 (4000) (block).

  Example of

  create table test (id primary number, data key blob);

  I want to display like

  Select code, data of the test;

  ID data

  1 character 4000

  1 character 4000

  1 character 4000

  .

  .

  and soon.

  Please help me on this

  BLOB stores binary data. What is the point of converted them to a varchar2? If the LOB will store the character data, then the appropriate data type would be CLOB. Said that, assuming that you will use CLOB here is the answer to your question.

  drop table temp purge;

  create table temp

  (

  integer ID

  clob c_lob

  );

  declare

  CLOB var;

  Start

  DBMS_LOB.CREATETEMPORARY (var, true);

  because me in 1.32 loop

  DBMS_LOB. WriteAppend (var, 1000, rpad ('x', 1000, 'x'));

  end loop;

  insert into temp (id, c_lob) values (1, var);

  insert into temp (id, c_lob) values (2, var);

  commit;

  end;

  /

  Select id

  rno

  , length (dbms_lob.substr (c_lob, 4000, offset)) c_lob_var_len

  , dbms_lob.substr (c_lob, 4000, offset) c_lob_var

  de)

  Select id

  , ((level-1) * 4000) + 1 offset

  c_lob

  level NWR

  temp

  connect

  by level<=>

  and prior id = id

  and dbms_random.value (prior) is not null

  );