SQL: REGEXP_SUBSTR CSV with string delimiter

Similar Questions

• You are looking for assistance on the use of CSV with InDesign files for merges data...

  Hello

  Please forgive me if this has been discussed in this forum already. If so, please point me to this discussion.

  What I'm trying to do, it's a data merge InDesign with a (CSV) file delimited by commas. I created the CSV file by using the command "SaveAs CSV" with Office Excel 2013. The problem I have is with records that contain multiple lines in the same area. After reviewing the merged InDesign document, I noticed that the results are inaccurate for the rest of this record. When it gets to the record following some fields are filled with the remaining records of the previous record.

  I'm sorry if I'm not explaining this very well, but the results are strange and I hope others have similar results and can offer a solution.

  Thank you

  Greg

  You can't have line breaks in a cell. The solution is to use a code for the break, and then run find/replace after the merger to change the code to break you.

• Chain of Split which begins with a delimiter

  Hi all
  I am trying to split a string using "-" as a separator and store values (strings) in a list. The problem is that, because the string begins with the delimiter, I get an empty (as my first item) element in my list. I would appreciate your help to solve this problem.
  
  My code is:
  

  public class SplitString { 
     public static void main(String[] args) {
       String str = "-Djava.security.policy=$ORACLE_HOME/j2ee/home/config/java2.policy" +
                    "-Djava.awt.headless=true -Dhttp.webdir.enable=false";
            
            SplitAndStoreAsList(str);
       }
       
      public static void SplitAndStoreAsList(String str){
       List<String> list = Arrays.asList(str.split("-"));
            
       for (int i = 0; i<list.size(); i++){
               System.out.println(i + ": " + "-" + list.get(i));
       }
            
      }
  }

  The output should be

  0: -Djava.security.policy=$ORACLE_HOME/j2ee/home/config/java2.policy
  1: -Djava.awt.headless=true 
  2: -Dhttp.webdir.enable=false

  But it is

  0: -
  1: -Djava.security.policy=$ORACLE_HOME/j2ee/home/config/java2.policy
  2: -Djava.awt.headless=true 
  3: -Dhttp.webdir.enable=false

  Why not just change the regular expression that is

       List list = Arrays.asList(str.split("(?
  

• Urgent need help with String Manipulation

  Hello;
  I really need help with strings.
  
  Let's say I have these channels:
  \\ABC\DEF\GHI\xyz\xyz\xyz\xyz
  
  \\ABC\DEF\GHI\xyz\
  
  \\ABC\DEF\GHI\xyz\xyz\xyz\
  
  \\ABC\DEF\GHI
  
  \\ABC\DEF\GHI\
  
  ABC/DEF/GHI/xyz/xyz/xyz/xyz
  
  ABC/DEF/GHI/xyz /.
  
  ABC/DEF/GHI/xyz/xyz/xyz /.
  
  ABC/DEF/GHI
  
  ABC/DEF/GHI.
  
  I need to convert them using a SQL statement for:
  \\ABC\DEF\GHI\xyz\ or //ABC/DEF/GHI/xyz/ according to the original management / or \. Please note that the "xyz" may be repeated zero or more times. I just need one end with a trailing / or -if she had one first.
  
  I'm on Oracle 11.1.6 SE. I have a bunch of attempts with instr, substr and replace, but I think that regexp_replace could be better.
  Any help would be GREAT! I have a whole team waiting on me.

  Hello.

  Here is a way.

  -----------------------------------------------------------------------
  with sample_table as
  (
       select '\\ABC\DEF\GHI\xyz\xyz\xyz\xyz' str from dual union all
       select '\\ABC\DEF\GHI\xyz\' from dual union all
       select '\\ABC\DEF\GHI\xyz\xyz\xyz\' from dual union all
       select '\\ABC\DEF\GHI' from dual union all
       select '\\ABC\DEF\GHI\' from dual union all
       select '//ABC/DEF/GHI/xyz/xyz/xyz/xyz' from dual union all
       select '//ABC/DEF/GHI/xyz/' from dual union all
       select '//ABC/DEF/GHI/xyz/xyz/xyz/' from dual union all
       select '//ABC/DEF/GHI' from dual union all
       select '//ABC/DEF/GHI/' from dual
  )
  select str,
       regexp_replace(str,'(xyz([/\])*)+$','\1') new_str
  from  sample_table;
  
  STR                      NEW_STR
  ------------------------------ ------------------------------
  \\ABC\DEF\GHI\xyz\xyz\xyz\xyz  \\ABC\DEF\GHI\xyz
  \\ABC\DEF\GHI\xyz\            \\ABC\DEF\GHI\xyz\
  \\ABC\DEF\GHI\xyz\xyz\xyz\     \\ABC\DEF\GHI\xyz\
  \\ABC\DEF\GHI                 \\ABC\DEF\GHI
  \\ABC\DEF\GHI\                 \\ABC\DEF\GHI\
  //ABC/DEF/GHI/xyz/xyz/xyz/xyz  //ABC/DEF/GHI/xyz
  //ABC/DEF/GHI/xyz/            //ABC/DEF/GHI/xyz/
  //ABC/DEF/GHI/xyz/xyz/xyz/     //ABC/DEF/GHI/xyz/
  //ABC/DEF/GHI                 //ABC/DEF/GHI
  //ABC/DEF/GHI/                 //ABC/DEF/GHI/
  

  Kind regards.

• Problem with string to Int conversion

  Dear friends,
  
  Problem with string to Int conversion
  
  I'll have a column where most of the values is digital. Only 4 values are not numeric.
  
  I have replaced the non-numeric values to numeric in order to maintain the data type.
  
  Grade.Grade DEAL WHEN "E4" THEN "24' WHEN 'E3' THEN '23' WHEN 'E2' THEN '22' WHEN 'E1' THEN '21' Grade.Grade OTHER END"
  
  This is the result as inactive
  
  Grade
  
  _0_
  
  _1_
  
  _10_
  
  _11_
  
  _12_
  
  _13_
  
  _14_
  
  15
  
  _16_
  
  _17_
  
  _18_
  
  _19_
  
  _2_
  
  _20_
  
  _21_
  
  _22_
  
  _23_
  
  _24_
  
  _3_
  
  _4_
  
  _5_
  
  _6_
  
  _7_
  
  _8_
  
  _9_
  
  Discount
  -            -            -
  
  Now I want to convert this value in digital and do some math
  
  So I changed the formula as below
  
  Cast (CASE Grade.Grade WHEN "E4" THEN "24' WHEN 'E3' THEN '23' WHEN 'E2' THEN '22' WHEN 'E1' THEN '21' Grade.Grade END as an INT. ELSE")
  
  
  Now I get the following error
  View display error
  Driver Odbc _ returned an error (SQLExecDirectW). _
  Error details
  Error codes: OPR4ONWY:U9IM8TAC:OI2DL65P
  _State: HY000. Code: 10058. [NQODBC] [SQL_STATE: HY000] [nQSError: 10058] A general error occurred. [nQSError: 17001] Oracle error code: 1722, message: ORA-01722: invalid number in the OIC calling OCIStmtFetch. [nQSError: 17012] In bulk fetch failed. _ (HY000)
  SQL issued: SELECT cast (CASE Grade.Grade WHEN "E4" THEN "24' WHEN 'E3' THEN '23' WHEN 'E2' THEN '22' WHEN 'E1' THEN '21' Grade.Grade END as an int. ELSE") seen'Human Capital' 0 - Costing Manpower WHERE LENGTH (Grade.Grade CASE WHEN 'E1' THEN '20' WHEN 'E2' THEN '21' WHEN 'E3' THEN '22' WHEN "E4" THEN "23 ' ELSE Grade.Grade END) > 0 ORDER BY saw_0_
  
  Discount
  
  
  
  
  
  Could someone help me
  
  Concerning
  
  Mustafa
  
  Published by: Musnet on June 29, 2010 05:42
  
  Published by: Musnet on June 29, 2010 06:48

  Are you sure you don't have another non-digital value in your column. I want to get this filter temporarily just to check if your plaster works or not. If it works, then you can remove this filter.

• string delimiter of pipe lines using regular expressions

  Can Hi anyone please help write pipe 'l' values of line splitter
  
  
  Samples "something1 |" something2 | something3'
  
  
  I tried with query below
  
  SELECT THE LEVEL,
  TRIM ("" ' OF RTRIM (REGEXP_SUBSTR ("something1 |")) "") something2 | something3' | ' |',
  '".*?"|| [^|] *|',
  1,
  LEVEL
  ),
  '|'
  )
  ) sub_str
  OF the double
  CONNECTION OF LEVEL < = LENGTH (REGEXP_REPLACE ("something1 |")) something2 | "something3 ','". * ? » | [^|] *')) + 1

  SQL> ed
  Wrote file afiedt.buf
  
    1  with t as (select 'something1|something2|something3' as txt from dual)
    2  -- end of sample data
    3  select REGEXP_SUBSTR (txt, '[^|]+', 1, level)
    4  from t
    5* connect by level <= length(regexp_replace(txt,'[^|]*'))+1
  SQL> /
  
  REGEXP_SUBSTR(TXT,'[^|]+',1,LEVE
  --------------------------------
  something1
  something2
  something3
  
  SQL>
  

• CSV with double quotes

  Hello

  I had tough time trying to figure out how to read a CSV with quotes, I need to read a csv file in order to fill out a worksheet.

  I am attaching a sample csv file to display the type of file, that I'm trying.

  Thank you

  After inspection, you're right, it's a trick question.  You need commas that follow an even number of quotes for separators.  Since my previous statement was not converted, I assume that you have an earlier version of LV, so I have attached an example of 8.6.  You can get rid of the quotes if you like (exercise for the reader).

  

• Reading a CSV with a large number of columns

  Hello

  I tried to read data from files large csv with 38 columns by reading a line using readline and scan using the scan linebuffer.

  The file size can be up to 100 MB.

  Scan does not seem to support the large number of areas.

  Suggestions to the comma separated fields 38 reading. There is a header row in the file.

  Thank you

  Have you considered the use of the modifier "rep"? You can find examples in help, search for 'chain with real table ASCII Comma-separated numbers '.

• want to use regexp_replace in pl/sql to replace a string containing repeat them consecutive to a words

  Hi Experts,

  I want to use regexp_replace in pl/sql to convert a string containing consecutive repeat words have only one.

  for example: "is my name and and I was waiting for waiting for what it" should be replaced by "is my name and I was waiting for it."

  Thanks in advance

  Gordon

  A slight change '(\w+\S\W)\1+' )

  Concerning

  Etbin

• java.sql.SQLException: problems with the native/lack of loading methods library

  Hi all
  
  Please let me know how to fix the exception "java.sql.SQLException: problems with the native/lack of loading methods library: no ttJdbc in java.library.path".
  
  Thank you
  Prabhu
  
  Published by: Nina Prabhu on November 20, 2012 02:12

  Hi Prabhu,

  Probably, you must specify the LD_LIBRARY_PATH variable. Like the following:

  export LD_LIBRARY_PATH=$TIMESTEN_HOME/ttoracle_home/instantclient_11_1
  

  Best regards
  Gennady

• Split a string using regexp_substr with consecutive delimiter

  I split a string with tubes as a delimiter. A string like this for example:

  "THIS |" EAST | ONE | EXAMPLE '

  If I do something like this:

  SELECT REGEXP_SUBSTR('THIS|IS|AN|EXAMPLE', '[^|]+', 1, 4) FROM DUAL
  

  I would get the word EXAMPLE

  But if the string is like this:

  "THIS |" EAST | ONE | EXAMPLE '

  With the above query, I always get EXAMPLE, but the word should be in the next position (5) because after IS, there should be an empty element

  Is it possible to change the regular expression to also get the empty element?

  Thanks in advance

  Well, the way I understand it, you get this behavior, because there is 'nothing' between the delimiters. so there is a field, so it does not display it don't think the regular expression.

  Try it with a space between the delimiters - works fine.

  So, with this in mind, the simplest solution might be something like that?

  regexp_substr (replace(c,'||','| |'), "[^ |]") +' 1, 4)

  [edited to avoid having a partial answer marked "correct"]

  As mentioned below by a few others, there are other solutions using regular expressions.

  The solution above does not work if more than 2 consecutive fields are empty, or if the first or last is empty.

  A simple tweak to the logic would help with this:

  RTrim (regexp_substr (replace (c,'|))) ',' |'), ' [^ |] +', 1, level))

  However, even if it has still some limitations compared to the solution of the full regular expression mentioned by Frank, below.

  [/ Edit]

• SQL Loader issue - CSV with commas and quotes IN the data

  Hello, I have a dataset for a simple table of 2 columns like this:

  Column 1, "it is given for"Column 2", with commas and quotes."

  Data are delimited by commas and may be surrounded by double quotes.  In ADDITION, it may include commas and quotation marks in the data fields.  I CANNOT manipulate data before sending it to SQL Loader.

  I set my file of control like this:

  DOWNLOAD THE DATA

  INFILE '. / TEST.dat'

  BADFILE '. / TEST. BAD'

  DISCARDFILE '. / TEST. DSC"

  REPLACE IN THE TEST TABLE

  Fields ended by ',' POSSIBLY BOX BY ' "' TRAILING NULLCOLS"

  (

  Col1 char (50),

  Col2 char (500)

  )

  
  

  Now when I run the present via SQLLDR, I get the following error in the log file:

  
  

  Sheet 1: Rejected - error on table TEST, column COL2.

  no terminator found after CLOSE and CLOSED field

  What are my options to get the loaded data as presented above?  I'm working on Oracle 11 g (11.2.0.3.0) 64-bit on AIX 6.1.0.0.

  Thank you!

  In this case, there is no way she can tell who's a delimiter or pregnant and which is part of the data.  As far as I know, there is no way that you can load it into the appropriate columns.

• Problem with regexp_substr if the string with null values between commas.

  Hi all
  
  My version of DB is 10.1.0.5.0
  I want to extract the string values using query below, but I am unable to provide the correct result.
  

  WITH t  AS ( select '123,1,3,22' col FROM DUAL
          UNION ALL
          SELECT '123,,2,1' FROM DUAL
          UNION ALL
          SELECT '5,1,2,,' FROM DUAL
          ) 
  SELECT regexp_substr(t.col,'[^,]*',1,1)
        ,regexp_substr(t.col,'[^,]*',1,2)
        ,regexp_substr(t.col,'[^,]*',1,3)
        ,regexp_substr(t.col,'[^,]*',1,4) FROM DUAL,t   ;

  My required result as below.

  123     1      3      22
  123            2       1
  5       1      2        

  Please help me to get the required result by using regular expressions.
  Thanks, Ram

  Hello

  In Oracle 10, you can do so by combining REGEXP_SUBSTR with something else like RTRIM:

  SELECT     RTRIM (REGEXP_SUBSTR (col, '[^,]*(,|$)', 1, 1), ',')     AS part_1
  ,     RTRIM (REGEXP_SUBSTR (col, '[^,]*(,|$)', 1, 2), ',')     AS part_2
  ,     RTRIM (REGEXP_SUBSTR (col, '[^,]*(,|$)', 1, 3), ',')     AS part_3
  ,     RTRIM (REGEXP_SUBSTR (col, '[^,]*(,|$)', 1, 4), ',')     AS part_4
  FROM     t
  ;
  

  Starting Oracle 11 REGEXP_SUBSTR alone can do.

  But look at the first response; regular expressions will be probably less effective than less powerful features, especially when all channels have the same number of separators.

  Published by: Frank Kulash on 29 December 2012 08:36

• REGEXP_SUBSTR for the list delimited by commas with null values

  Hello
  
  I have a column that stores a list of values comma-delimited. Some of these values in the list may be null. I'm having some trouble trying to extract the values using the REGEXP_SUBSTR function when null values are present. Here are two things I've tried:

  SELECT
     REGEXP_SUBSTR (val, '[^,]*', 1, 1) pos1
    ,REGEXP_SUBSTR (val, '[^,]*', 1, 2) pos2
    ,REGEXP_SUBSTR (val, '[^,]*', 1, 3) pos3
    ,REGEXP_SUBSTR (val, '[^,]*', 1, 4) pos4
    ,REGEXP_SUBSTR (val, '[^,]*', 1, 5) pos5
  FROM (SELECT 'AAA,BBB,,DDD,,FFF' val FROM dual);
  
  POS P POS P P
  --- - --- - -
  AAA   BBB

  SELECT
     REGEXP_SUBSTR (val, '[^,]+', 1, 1) pos1
    ,REGEXP_SUBSTR (val, '[^,]+', 1, 2) pos2
    ,REGEXP_SUBSTR (val, '[^,]+', 1, 3) pos3
    ,REGEXP_SUBSTR (val, '[^,]+', 1, 4) pos4
    ,REGEXP_SUBSTR (val, '[^,]+', 1, 5) pos5
  FROM (SELECT 'AAA,BBB,,DDD,,FFF' val FROM dual);
  
  POS POS POS POS P
  --- --- --- --- -
  AAA BBB DDD FFF

  As you can see that neither calls work correctly. Anyone know how to change the regular expression pattern to handle null values? I tried various other models but could not get anyone to work in all cases.
  
  Thank you
  
  Martin
  -----
  http://www.ClariFit.com
  http://www.TalkApex.com

  Hi, Martin,.

  That's what you want:

  SELECT
     RTRIM (REGEXP_SUBSTR (val, '[^,]*,', 1, 1), ',') pos1
    ,RTRIM (REGEXP_SUBSTR (val, '[^,]*,', 1, 2), ',') pos2
    ,RTRIM (REGEXP_SUBSTR (val, '[^,]*,', 1, 3), ',') pos3
    ,RTRIM (REGEXP_SUBSTR (val, '[^,]*,', 1, 4), ',') pos4
    ,RTRIM (REGEXP_SUBSTR (val || ','
                        , '[^,]*,', 1, 5), ',') pos5
  FROM (SELECT 'AAA,BBB,,DDD,,FFF' val FROM dual);
  

  The query above works in Oracle 10 or 11, but in Oracle 11, you can also do it with only REGEXP_SUBSTR, without using RTRIM:

  SELECT
     REGEXP_SUBSTR (val, '([^,]*),|$', 1, 1, NULL, 1) pos1
    ,REGEXP_SUBSTR (val, '([^,]*),|$', 1, 2, NULL, 1) pos2
    ,REGEXP_SUBSTR (val, '([^,]*),|$', 1, 3, NULL, 1) pos3
    ,REGEXP_SUBSTR (val, '([^,]*),|$', 1, 4, NULL, 1) pos4
    ,REGEXP_SUBSTR (val, '([^,]*),|$', 1, 5, NULL, 1) pos5
  FROM (SELECT 'AAA,BBB,,DDD,,FFF' val FROM dual);
  

  The problem with your first request was that he was looking for sub channels of 0 or more non-virgules. There was as a substring. consisting of 3 characters starting at position 1, he returned "AAA", as expected. Then there was an another substring, the 0 characters, starting at position 4, so it returned NULL. Then, there was a substring of 3 characters starting at position 5, so he returned 'BBB '.

  The problem with your 2nd request was that he was looking for 1 or more non-virgules. 'DDD' is the 3rd this substring.

  Published by: Frank Kulash, on February 16, 2012 11:36
  Added Oracle 11 example

• How to divide a column (row) with a '&amp;' delimited string by several lines

  Hello
  
  I need to divide a column of data in several rows as shown below. Could someone help me solve this problem
  
  CREATE TABLE TEMP_ASS_BOOK
  (
  book_id NUMBER,
  BOOK_TYPE VARCHAR2 (50 BYTE),
  BOOK_NAME VARCHAR2 (50 BYTE),
  ASS_BOOK VARCHAR2 (100 BYTE)
  )
  
  and data as below:
  
  Insert into TEMP_ASS_BOOK
  (BOOK_ID, BOOK_TYPE, BOOK_NAME, ASS_BOOK)
  Values
  (1, SCIENCE, 'PHYSICAL INTER', 'JP & THISTLE & SPECTRUM SERIES SERIES SERIES');
  Insert into TEMP_ASS_BOOK
  (BOOK_ID, BOOK_TYPE, BOOK_NAME, ASS_BOOK)
  Values
  (2, MATH, "MATH INTER', 'RK SERIES SERIES SPECTRUM & GVK SERIES');
  
  
  
  Now, I need to insert this data into another table as below:
  
  CREATE TABLE ASS_BOOK
  (
  book_id NUMBER,
  BOOK_TYPE VARCHAR2 (50 BYTE),
  BOOK_NAME VARCHAR2 (50 BYTE),
  ASS_BOOK VARCHAR2 (100 BYTE)
  )
  
  now insert data from TEMP_ASS_BOOK to ASS_BOOK as below:
  
  1 1 SCIENCE PHYSICS INTER SERIES JP
  2 1 SCIENCE PHYSICS INTER SERIES THISTLE
  3 1 INTER SPECTRUM PHYSICAL SCIENCE SERIES
  4. 2 MATH MATH INTER SERIES RK
  5 2 MATH MATH INTER SPECTRUM SERIES
  6 2 MATH MATH INTER GVK SERIES
  
  Thanks in advance.

  I agree with knani, it is not necessary to resort to a procedure...

  However, if you do not use the PL/SQL easier and more understandable code to do the job:

  declare
  cursor books is select * from temp_x_book;
  begin
    for b in books loop
       for i in 1..length(b.x_book)-length(replace(b.x_book,'&'))+1 loop
           insert into x_book (BOOK_ID,BOOK_TYPE,BOOK_NAME,X_BOOK)
                       values (b.book_id,b.book_type,b.book_name,regexp_substr(b.x_book, '[^&]+', 1, i));
       end loop;
    end loop;
  end;
  /
  

  Max
  [My Italian blog Oracle | http://oracleitalia.wordpress.com/2009/12/27/inviare-email-dal-db-utilizzando-utl_smtp/]