RegExp as

Similar Questions

• standard character of regexp

  Hello

  I'm looking for a regexp avoid the user using anything else than 1-9 a - z A - Z and '_' to a file name.

  concerning

  Tinnitus

  [^ \w] matches anything that is not an alphabetic character.

• Limiting the tables and using lindex in commands regexp

  Hello

  Here's my second and extremely crap question/post in what concerns the EEM/TCL.  I did some research on Google and have not really been able to find an answer to these questions.  Probably more due to my lack of agility with the search bar of the information being just is not there.  Hope you'll forgive me if.

  1. my first question is this.  Say I want to get an output of a command.  I would like to see all the interfaces in the 'ip vrf command show interface' for example.  Not knowing how many interfaces there may be, there could be 10, or there might be 200, what is the best way to limit a loop function after that I gave the following commands?

  Set _vrf_int [exec show ip vrf interface]
  Set _array_vrf_int [split $_vrf_int '\n']

  set $i 0

  then {$i< 200}="">
  RegExp {([A-Z]+[a-z]+[0-9]+). *} [lindex $_array_vrf_int $i] _complete_string _int
  Inc. $i 1

  }

  Not 100% sure that the syntax is correct, sorry.  If I don't have that 10 interfaces so it won't be necessary to issue this loop 200 times to "$_array_vrf_int".  Also wouldn't be great if I went through the 200 interfaces.  Is there a way I can limit this loop based on the number of lines that were captured after the split function?

  2. the value of the regular expression above is output in the var $_int.  Is there a way I can output the regular expression in a table? a ' :-

  RegExp {([A-Z]+[a-z]+[0-9]+). *} [lindex $_array_vrf_int $i] _complete_string [lindex $ $i _int]

  Syntax is certainly not correct, but I guess there could be a way to do this.  Any thoughts?

  3. I did some comparisons on an outing to see if an access list is present on an interface.  I can get a regexp to pull back the name of the access on a given interface list, but if there are any access-list to an if statement on anything?

  Set _acl_name 0

  RegExp {. * ip access-group (. *) in} $_running_int _complete_string _acl_name

  If {$_acl_name! = 0} {puts ' there is an access-list on $_int "} else {puts ' there is no acl on _int $ :-("} "}

  In the above output I'm defining the value of ' $_acl_name ' to zero so that I can compare it to the fi statement.  This seems to be a newbie to shit how do however.  Is it a kind of generic I put in place of the '0' to match an empty variable. has ' {null}.

  4 I'm sorry last question.  I also reset the variable {null} using a similar wildcard.  Once again I can reset the variable to 0, but it seems just that I do not understand that the syntax well enough and it is without doubt a better method.

  5. I have read several tutorials that cover a lot of bases.  Is there a good reference that anyone can suggest so I don't waste your time with these silly questions?  I'm afraid that the scripts I've studied on this forum are always way above my head.

  Thanks in advance

  Alex

  Assume that the limit of the loop.  Which will never end well.  Instead, use a foreach loop to iterate over the number of lines in the output:

  set vrf_int [exec "show ip vrf interface"]
  foreach line [split $vrf_int "\n"] {
  
        if { [regexp {([A-Za-z0-9]+).*} $line -> int_name] } {
  puts "Interface name is $int_name"
  }
  }
  

  In addition, do not use the variables that start with "_".  Those that are reserved for the Cisco or overall use.  In addition, 'exec' is a single command tclsh.  If you use the EEM Tcl, you will need to interact with the CLI library.  You can watch our best practices guide at https://supportforums.cisco.com/docs/DOC-12757 for some tips and tricks with the use of EEM.

  You can use - all and--inline for regexp all return in a list.  However, given the idea of the loop above, you can also use lappend to each pass:

  set intlst [regexp -all -inline {([A-Za-z0-9]+).*} $output]

  
  set vrf_int [exec "show ip vrf interface"]
  set intlst [list]
  foreach line [split $vrf_int "\n"] {
  
        if { [regexp {([A-Za-z0-9]+).*} $line -> int_name] } {
  lappend intlst $int_name
  }
  }
  
  
  
  You can use "info exist" or simply to check the result of your regexp command to see if a match took place.  See the example above the latter.
  regexp {.*ip access-group (.*) in} $_running_int _complete_string _acl_name
  if {[info exists $_acl_name]} {puts "there is a access-list on $_int"} else {puts "there's no acl on $_int :-("}
  
  You can 'reset' a variable by using the command 'deactivated '.
  One of the best general Tcl references is the "book of the pen:
  
  http://www.Amazon.com/practical-programming-Tcl-4th-Edition/DP/0130385603/ref=sr_1_2?ie=UTF8&QID=1333817448&SR=8-2
  		   

• RegExp

  Hi all

  To exclude it using the regexp, if I want to "has" the value of the name column 'a' need to remove data from remaning.

  Insert into TEST (NAME) Values ('a, b, c');

  Required: if I spend 'a'-> 'b, c' should be deleted-> output-> in the name column

  Required: if I switch "b"-> 'a, c' should be deleted-> output-> b in the name column

  Thank you

  Expect update the value you pass? Under update check if the value passed exists or not... If there is, it will update this line with whatever it is you've spent...

  -Updated (pname is the setting which you will pass)

  UPDATE test_1 SET name =: pname

  OÙ «, » || name | ',' LIKE '%', | : pname |', %';

  -Test cases

  -Passing the value of the parameter in the form "a".

  SQL > pname VARIABLE VARCHAR2 (1);

  SQL > EXEC: pname: = 'a ';

  PL/SQL procedure successfully completed.

  -Before the update

  SQL > SELECT * FROM test_1;

  NAME

  ----------

  a, b, c

  d, e, f

  -Updated the line given by "a".

  SQL > UPDATE test_1 SET name =: pname

  2. WHERE «,» | name | ',' LIKE '%', | : pname |', %';

  1 line update.

  -After the update

  SQL > SELECT * FROM test_1;

  NAME

  ----------

  one

  d, e, f

  SQL > ROLLBACK;

  Complete restoration.

  SQL > SELECT * FROM test_1;

  NAME

  ----------

  a, b, c

  d, e, f

  -Value of the parameter as 'b '.

  SQL > EXEC: pname: = 'b ';.

  PL/SQL procedure successfully completed.

  SQL > UPDATE test_1 SET name =: pname

  2. WHERE «,» | name | ',' LIKE '%', | : pname |', %';

  1 line update.

  SQL > SELECT * FROM test_1;

  NAME

  ----------

  b

  d, e, f

  SQL > ROLLBACK;

  Complete restoration.

  SQL > SELECT * FROM test_1;

  NAME

  ----------

  a, b, c

  d, e, f

  -Value of the parameter as 'c '.

  SQL > EXEC: pname: = 'c ';

  PL/SQL procedure successfully completed.

  SQL > UPDATE test_1 SET name =: pname

  2. WHERE «,» | name | ',' LIKE '%', | : pname |', %';

  1 line update.

  SQL > SELECT * FROM test_1;

  NAME

  ----------

  c

  d, e, f

• Customized using regexp check constraint

  Oracle 11.2.0.1

  I need a customized using regexp expressions check constraint.  We have a single column in our table REGNO, which is used to keep the registration number of our customers.  It has alpha and numeric values, the first is Alpha and rest is digital with preceding zeros as A0100 A1245, C1111 etc.  Sometimes, in the application user enters regnos evil if we want to stop entering the bad regnos i.e. assume that the user name is Ashok Kumar and in the series, we have 1000 users so this new user should have A1001 (other than any value A1001, in this example, the constraint must be limit to enter the value) only, not the others nor a1001 , A1000, nor A1002 or < has no > 1001 too. Rather than having 26 sequences (A to Z), we get a table where we have these columns:

  Alpha char (1), number of LastNo.  But sometimes a user enters wrong number value and stores it in the table, so next time for the same gap alpha arrives and confusing.

  So, we must have a constraint of database level so that ONLY the correct values must be entered.

  Kindly help me how I have this task please.

  Thank you.

  If REGNO is a derived value then why let you the user enter a value. Just generate the user.

• RegExp search and replace, keep the formatting of the original text

  Hello!

  Will have to perform the Search and replace within a framework of text using RegExp.

  And to retain the original formatting of the text block.

  How is it possible?

  Bad example-, it converts all the text formatting to the format of the first character:

  var reg = /a/gmi;
  var replacer = '*';
  var fr = activeDocument.textFrames[0];
  fr.contents = fr.contents.replace (reg , replacer);
  

  

  Thank you!

  I has not yet been tested your code.

  But why you did not play with my code snippet, use something like this:

  // regex_changeContentsOfWordOrString_RemainFormatting.jsx
  // regards pixxxel schubser
  var s = /arguments/gi;
  var replacer = "other string", result;
  var atf = activeDocument.textFrames[0];
  
  while (result = s.exec(atf.contents)) {
      try {
          aCon = atf.characters[result.index];
          aCon.length = result[0].length;
          aCon.contents = replacer;
          } catch (e) {};
      }
  

  Try it and have fun

• regexp help

  Hello

  I need to extract deposited number. Can please help me this extract using regexp.

  for ex:

  at the bottom of the message, I need to extract the value 434234234235435345634345 using REGEXP

  ORA-20102: cw_upload_queue_item_meta | ec3_unique_file_ref:434234234235435345634345 | ORA-00001: unique constraint (IBIS. CW_UPLOAD_QUEUE_ITEM_META_PK) violated

  user575115 wrote:

  Thx.It works in 11g, 10g too error throw

  That's why you have to mention your db version 4-digit during the validation of the questions:

  Select

  LTRIM (regexp_substr ('ORA-20102: cw_upload_queue_item_meta | ec3_unique_file_ref:434234234235435345634345 |)) ORA-00001: unique constraint (IBIS. ((CW_UPLOAD_QUEUE_ITEM_META_PK) violated ',': [^ |] +', 1, 2),': ')

  f

  of the double

• IAM facing problems sql query regexp

  Hi all

  under custom query note here is the type of clob data in table

  value of the field personalized note is exists table Anshu Udainiya the value tag iam using the regular expression to replace

  Select org_prog_cam_id, ndc_no, REGEXP_REPLACE (trim (to_char (custom_note)),' < [^ <>] + > ') custom_note

  from DTC_AAG_NDC_NOTE_STAGING where org_prog_cam_id = 6696 and ndc_no = '51248015003'

  less

  Select 6696, '51248015003', to_char ('sdfsfsdfsd'))

  of the double

  first query of exit

  ndc_no	cust1	cust2
  51248015003	< p > < span style = "color: #00FFFF" > < strong > < span style = "font-size: 16px" > < span style = "background-color: #FFD700" > Anshu Udainiya </span > < / span > < / strong > < / span > < /p >	16

  second length of request of personalized note

  Choose the length (REGEXP_REPLACE (trim (to_char ('Anshu unmar')),' < [^ <>] + > ')) coz of the double

  coz

  -----

  14

  can you please help these for urgent need... Thank you very much

  Hello

  assuming that it is the only element between 2 beacons (between > and)<) ans="" that="" you="" are="" using="" at="" least="" 11g="" you="" can="" use="" regexp="" in="" this="">

  with t
  as
  (
  select '
  Anshu Udainiya
  ' str
    from dual
  )
  select regexp_substr(str, '>([^<]+)<',1,1,null,1) txt, length(regexp_substr(str, '>([^<]+)<',1,1,null,1)) len
    from t;
  
  TXT                   LEN
  -------------- ----------
  Anshu Udainiya         14
  

  Kind regards.

  Alberto

• RegExp in a case of join internal

  Hi, I was expecting assistance regarding tent to match values using regexp because currently I seem to get an error. I created a set of dummy tables to give an idea of the problem that I am facing and here is the table create and insert the codes.

  
  CREATE TABLE "BAS_USERS" 
     ( "UID" VARCHAR2(20 BYTE), 
   "FIRSTNAME" VARCHAR2(255 BYTE), 
   "LASTNAME" VARCHAR2(255 BYTE)
     )
   
  Insert into BAS_USERS ("UID",FIRSTNAME,LASTNAME) values ('123456789','Humpty','D');
  Insert into BAS_USERS ("UID",FIRSTNAME,LASTNAME) values ('987654321','Mickey','M');
  Insert into BAS_USERS ("UID",FIRSTNAME,LASTNAME) values ('543212345','Minnie','M');
  
     
  CREATE TABLE "BAS_USERS_LOG" 
     ( "NOTES" VARCHAR2(255 BYTE)
     )
  
  
  Insert into BAS_USERS_LOG (NOTES) values ('TASK WAS ASSIGNED BY Humpty D (123456789) TO Mickey M (987654321) ');
  Insert into BAS_USERS_LOG (NOTES) values ('TASK WAS FETCHED BY Minnie M (543212345) ');
  
   
  

  What I'm trying to do is to try to take the name of the value of the field notes in my table of bas_users_log and try then to match it by using the CONCATENATE function to combine the first and last name, but for some reason that I get the following error:

  ORA-00920: invalid relational operator
  00920. 00000 -  "invalid relational operator"
  *Cause:    
  *Action:
  Error at Line: 12 Column: 5
  

  The SQL Code I wrote is this:

  
  SELECT
      *
  FROM
      BAS_USERS_LOG B
  INNER JOIN 
      BAS_USERS D
  ON
      CASE
          WHEN (REGEXP_SUBSTR(REGEXP_REPLACE(UPPER(TRIM(B.NOTES)),'-',''), '[A-Z]+', 1, 8) ||' '|| REGEXP_SUBSTR(REGEXP_REPLACE(UPPER(TRIM(B.NOTES)),'-',''), '[A-Z]+', 1, 9))=' ' 
          THEN (REGEXP_SUBSTR(REGEXP_REPLACE(UPPER(TRIM(b.NOTES)),'-',''), '[A-Z]+', 1, 5) ||' '|| REGEXP_SUBSTR(REGEXP_REPLACE(UPPER(TRIM(B.NOTES)),'-',''), '[A-Z]+', 1, 6))
          ELSE (REGEXP_SUBSTR(REGEXP_REPLACE(UPPER(TRIM(B.NOTES)),'-',''), '[A-Z]+', 1, 8) ||' '|| REGEXP_SUBSTR(REGEXP_REPLACE(UPPER(TRIM(B.NOTES)),'-',''), '[A-Z]+', 1, 9)) 
  END AS B.PERSONNAME =(trim(UPPER(D.FIRSTNAME))|| ' ' ||trim(UPPER(D.LASTNAME)))
  ;
  

  Would appreciate it if someone could please advise where I'm wrong.

  Thanks in advance.

  just get rid of the column alias in your join condition (AS B.PERSONNAME)

  HTH

• REGEXP: String normalization

  Hi all

  Is it possible to normalize a string using a REGEXP to follow this rule:

  
  1. If the string begins with BLOCO or LOTE channels, keep it in the beginning of the string.
  2. keep the first channel number
  3 - If the first number is followed (after a space) that a single character include the number
  4 - If the first number is concatenated characters keep the whole word.

  
  
  Examples of data with examples:

  WITH t AS
  (SELECT 'LOTE 1' input_string,
           'LOTE 1' output_string
      FROM dual
    UNION ALL
    SELECT 'BLABLA 1' input_string,
           '1' output_string
      FROM dual
    UNION ALL
    SELECT 'BLOCO 11 A 3' input_string,
           'BLOCO 11A' output_string
      FROM dual
    UNION ALL
    SELECT 'BLOCO 3A' input_string,
           'BLOCO 3A' output_string
      FROM dual
    UNION ALL
    SELECT '4 A' input_string,
           '4A' output_string
      FROM dual
    UNION ALL
    SELECT '5 AA' input_string,
           '5' output_string
      FROM dual
    UNION ALL
    SELECT 'LOTE 6 RC' input_string,
           'LOTE 6' output_string
      FROM dual
    UNION ALL
    SELECT 'LOTE 7-A' input_string,
           'LOTE 7-A' output_string
      FROM dual)
  SELECT *
    FROM t;
  
  

  Kind regards

  Manel

  Post edited by: Manuel Vidigal Typo in "BLOCO 11 A 3"output ".

  Then

  Select

  input_string

  output_string

  regexp_substr)

  regexp_replace (input_string, ' (\d) ([^ [: digit:]] (| $))', "\1\2", 1, 1)

  ,'(^ BLOCO. | ^Lote) [^ [: digit:]] * \d [^] * | \d[^] * "

  ) o

  t

  Some explanantion:

  First in the regexp_replace replace us (I modified it) the first occurrence of a followed by a space and a number digt not) with numbers and his successor non-chiffres.

  And then from there we take if the string begins with BLOCO or LOTE everything from the beginning selection until the first digit and after this figure in the next space or at the end if there is no following space.

  If there is no key words at the beginning only since the first digit comes from the same way as above.

• RegExp is 10 x slower in Adobe Air

  This code run more slowly in Adobe Air 10 x:

  import flash.utils.getTimer;

  var str:String = "I'm a short string, I am a short string, I am a short string, I am a short string, I am a short string";

  var r: RegExp = /rt | tr | m | a | n/g

  var t:int = getTimer();

  var: string result;

  for (var i: uint = 0, n: uint = 100000; i < n; ++ i)

  {

  result = str.replace (r, ' |');

  }

  trace (getTimer () - t);

  trace (result);

  I run this code in Flash Player and Adobe Air (publication Mode, win 10 64-Bit, Dell Inspiron 3542) and get the result:

  Flash Player: 874 ms

  Adobe Air: 5461 ms

  I tested this code with CC to animate and develop Flash and get the same result. Sometimes I 10 x slower.

  I posted this problem here: Bug #4161031 - regular expressions (RegExp) is 10 x slower in Adobe Air

  This is a strange question. All this interest please do the test and report the result here?

  Thank you!

  Hello

  AIR SDK 22 looks good, can you use it?

  -Roshan

• How to search for a string in a column using REGEXP

  
  Hi Forum,

  I have a table with a column that has records that contain one of the 3 words viz.

  TABLE X

  col1

  -----------------------------------------------------------------------------

  remit_method = EMAIL/remit_details = some other details

  remit_method = SMS/remit_details = some other details for sms

  remit_method = POST

  I need to write a SQL that captures the remit_method of the column. I wrote SQL using SUBSTR and INSTR to capture him. But I was wondering if I could do it using the functions of RegExp as well?

  SELECT REGEXP_SUBSTR (col1) == > how?

  Of

  (

  SELECT ' remit_method = EMAIL/remit_details = FROM DUAL some other details

  UNION

  SELECT ' remit_method = SMS/remit_details = some other details for FROM DUAL of sms

  UNION

  SELECT 'remit_method = POST' FROM DUAL

  ) ;

  
  

  Output must be

  1st row = > EMAIL

  Row2 = > SMS

  Row3 = > POST

  Thank you

  Maëlle

  Like this...

  SQL > ed
  A written file afiedt.buf

  1 with tablex as (SELECT ' remit_method = EMAIL/remit_details = col1 OF DOUBLE UNION some other details)
  2. SELECT ' remit_method = SMS/remit_details = some other details OF UNION of the sms for DOUBLE
  3. SELECT 'remit_method = POST' FROM DUAL
  4                 )
  5 SELECT REGEXP_REPLACE (col1, ' ^ remit_method =([[:alpha:]]+). * $', '\1') as remit_method
  REGEXP_REPLACE 6, (col1, ' ^. * remit_details =(.*) $ |. *', '\1') as remit_details
  7 * FROM tablex
  SQL > /.

  REMIT_METHOD REMIT_DETAILS
  -------------------- ----------------------------------------
  A few other details by E-MAIL
  Publish
  Other details for sms SMS

• REGEXP question

  I'm trying to match, verify via REGEXP_LIKE:

  "start - something - dot - something - point - something - dot - something - end dot - digit numbers [---]-" / / or described differently *. * .digits [---] *. * .digits

  So for example the following must "match" 'I2.2.2[-]3.4.2', ' 1.1.3 1.1.4 of the[-]','1.2AA.3 [---] 1.2AB.4'

  but does not correspond to the following: "[-] 1.3 1.4 '.

  Somehow, I got lost in all the attempts to do right and ended up upward with solutions or scenarios or no match at all for the...

  The following seems to work as long as I do not add the ^ or $ signs... Someone at - it an idea?

  Select the check BOX

  WHAT model = 'X' AND 'FIT '.

  WHAT model <>'X' THEN 'NO MATCH'

  ANOTHER dummy

  End UP like "IS this IS IT? ''

  of the double

  where REGEXP_LIKE ('I2.2.4 [-] 3.2.44',' (\d+|\D+)\. () \d+|\D+)\.\d+\[\-\](\d+|\D+)\. (\d+|\D+)\.\d+');

  Version of database 11.2.0.3 EE

  Thank you.

  with t as)

  Select 'II.17B.2 [---] II.17B.4' double str

  Union of all the

  Select ' 3.2 [---] 3.4 ' double

  )

  Select

  *

  t

  where REGEXP_LIKE (str

  --                 ,'^(\d+?| \D+?*\.) {2}\d+?\[\-\]\d+?| \D+?*\.\d+?| \D+?*\.\d+?$'

  ,'^([^.] +\.) {2}\d+\[-\]([^.] +\.) \d+$ {2}'

  );

  Your regexp invites me to analyze it, so I give you only one solution.

  Just try to implement your needs forward and avoid things like. * ? If possible.

  During the construction, it is often useful to consider negative expressions regexp characterclasses.

  If your condition can be rewritten in the form

  'end of-no dots - dot - non-dots - dot - figures of the non-dots - dot - non-dots - dot - figures - [---] to start' / / or described differently *. * .digits [---] *. * .digits

  Well, I have to admit that the hypothesis "something = non-points" is a guess, but if it is a problem we can relax the cardinalities with {2}.

• RegExp associated

  
  Hello

  I'm new to regular expressions and I learn the techniques, can help me by telling more about branch reset the group in regex with some examples.

  From what I can understand of their article, it's something like that...

  Oracle: -.

  (?| ABC | ((d) (e) (f) | g (h) i)
  |      \_/\_/\_/  \_/
  |       |  |  |    |
  |       |  |  |    \--backreference 5
  |       |  |  |
  |       |  |  \--backreference 4
  |       |  |
  |       |  \--backreference 3
  |       |
  |       \--backreference 2
  |
  |
  \--------------------/
  |
  backreference 1

  Perl with branch Reset group: -.

  (?| ABC | ((d) (e) (f) | g (h) i)
  \_/\_/\_/  \_/
  |  |  |    |
  |  |  |    \--backreference 1
  |  |  |
  |  |  \--backreference 3
  |  |
  |  \--backreference 2
  |
  \--backreference 1

  Thus, the difference with the direction General Reset group is that each condition "OR" in the expression returns the backreferences 1, matches any group of GOLD allows the first element of backreferenced within this group to be named backreference 1, the second backreference as 2 and so on.

  In existing features of regexp in Oracle, it is one of the common problems (and probably why this feature has been introduced, but not in 11g as far as I can tell), each hook opening just increments the backreference, regardless of what group of GOLD, there is within the.  Which can make it very difficult to have a singular expression or matches, where you can reference the correct backreferences when you for example a regexp_replace.   The other problem with the current functionality is that the backreferences are limited of numbers from 1 to 9, so often we end up resulting using several RegExp functions to achieve what we want.  This new feature will definitely help (if oracle implement) as back-references are better interpreted and delivered to zero for each group.

• Need help with the regexp

  I am trying to run the query in oarcle apps and wanted to select lines, that meets the multiple org code separated by and my input value is 127 123, it will convert to ('127 ', ' 123') this regexp

  Select * from ORG_ORGANIZATION_DEFINITIONS where organization_code IN (select ' ('| regexp_replace (regexp_replace (regexp_replace (: p_inv_org_code, '() + ',','), ',' "'," '),'^ | $', "') |')) (' double)

  -NGO

  Hello

  This is the forum for the SQL Developer tool. Any questions about SQL or PL/SQL problems please post in SQL and PL/SQL

  Concerning

  Marcus

• RegExp with padding for the numbers

  Hi all

  I have a column VARCHAR2 representing Article and paragraph of a document and I need to sort in a sophisticated this column method.

  The problem is that because this value is from a form using free text, the format is quite heterogeneous. Then we would have something like:

  select '1' as section from dual union all
  select '7(1), 8(3)' as section from dual union all
  select '9(2)(b)' as section from dual union all
  select '11(1)(c)' as section from dual union all
  select 'Annex VI' as section from dual;
  

  the problem is that if we use the lexicographical order, that we'd end up with 11 front of 9.2 b (1)

  select  section from (
  select '1' as section from dual union all
  select '7(1), 8(3)' as section from dual union all
  select '9(2)(b)' as section from dual union all
  select '11(1)(c)' as section from dual union all
  select 'Annex VI' as section from dual)
  order by 1;
  
  SECTION
  ----------
  1        
  11(1)(c) 
  7(1), 8(3)
  9(2)(b)  
  Annex VI 
  

  It is true that articles go not beyond say 1000 thought I replace all the groups of digits of 0 left padded values. An expression that I would get the right result

  select  section from (
  select '001' as section from dual union all
  select '007(001), 008(003)' as section from dual union all
  select '009(002)(b)' as section from dual union all
  select '011(01)(c)' as section from dual union all
  select 'Annex VI' as section from dual)
  order by 1;
  

  SECTION

  ------------------

  001

  007 (001), 008 (003)

  009 (002) (b)
  011 (01) (c)
  Annex VI

  This can be achieved with regexp?, the thing is I can locate the numbers and normally the depth of numbers would be only 2 directions 1 (2) and no 1 (2) (3) so with 2 nested regexp_replace could be good. The thing is that I don't realize to fill in the 3rd parameter:

  Best regards

  Hello

  Using REGEXP_REPLACE to fill in the numbers:

  SELECT section

  , REGEXP_REPLACE (REGEXP_REPLACE (section - for debugging only

  , '(\d+)'

  , "000\1".

  )

  , « 0+(\d{4}) »

  , '\1'

  ) AS sort_str

  ARTICLE

  ORDER OF REGEXP_REPLACE (REGEXP_REPLACE (section

  , '(\d+)'

  , "000\1".

  )

  , « 0+(\d{4}) »

  , '\1'

  )

  ;

  The inner REGEXP_REPLACE adds 3 major 0 ' to all numbers, so that all the numbers are at least 4 digits.

  Removes from the external REGEXP_REPLACE 0 ' to all numbers, so that all numbers are exactly 4 digits.

  In Production, you will not want to repeat the expression in the SELECT clause. I just did here to help us see what's going on.

  Output:

  SECTION SORT_STR

  ---------- ------------------------------

  1 0001

  7 (1), 8 (3) 0007 (0001) 0008 (0003)

  9 2 b 0009 (0002) (b)

  11.1 c 0011 (0001) (c)

  Annex VI to annex VI