I have an ETL task for datawarehouse-ing purposes, I need to extract the second part of a String after a delimiter occurence such as: '#', 'ý', '-'. For example test case string:
'Tori 1#MHK-MahallaKingaveKD' I should retrieve only 'MHK'
'HPHelm2ýFFS-Tredddline' I should retrieve only 'FFS'
I already tried using the cases above:
TRIM(CASE
WHEN INSTR('HPHelm2ýFFS-Tredddline', '#',1,1) > 0
THEN (REPLACE(
REGEXP_SUBSTR('HPHelm2ýFFS-Tredddline', '[^#]+', 1,2),
'#'
))
ELSE (CASE
WHEN INSTR('HPHelm2ýFFS-Tredddline', '-',1,1) > 0
THEN (REPLACE(
REGEXP_SUBSTR('HPHelm2ýFFS-Tredddline', '[^-]+', 1,2),
'-'
))
ELSE (CASE
WHEN INSTR('HPHelm2ýFFS-Tredddline','-') = 0 AND INSTR('HPHelm2ýFFS-Tredddline','ý') = 0 AND INSTR('HPHelm2ýFFS-Tredddline','#') = 0
THEN 'HPHelm2ýFFS-Tredddline'
ELSE (CASE
WHEN INSTR('HPHelm2ýFFS-Tredddline','ý',1,1) > 0
THEN (REPLACE(
REGEXP_SUBSTR('HPHelm2ýFFS-Tredddline', '[^ý]+', 1,2),
'ý'
))
END)
END)
END)
END)
Using the code above I can retrieve:
'Tori 1#MHK-MahallaKingaveKD' ====> 'MHK-MahallaKingaveKD'
'HPHelm2ýFFS-Tredddline' ====> 'FFS-Tredddline'
Expected output:
'Tori 1#MHK-MahallaKingaveKD' ====> 'MHK'
'HPHelm2ýFFS-Tredddline' ====> 'FFS'
So I have to exclude '-' and the string after.
I guess I should modify the regexp_substr pattern but can't seem to find a clear solution since '-' is specified in the case when statements as a delimiter.
I suggest retrieving the second occurrence of 1+ chars other than your delimiter chars:
regexp_substr(col, '[^#ý-]+', 1, 2)
Here, the search starts with the first char in the record (1), and the second occurrence is returned (2).
The [^#ý-]+ pattern matches one or more (+) chars other than #, ý and -.
The following will give you what you're looking for:
WITH cteData AS (SELECT 'Tori 1#MHK-MahallaKingaveKD' AS STRING FROM DUAL UNION ALL
SELECT 'HPHelm2ýFFS-Tredddline' FROM DUAL)
SELECT STRING, REGEXP_SUBSTR(STRING, '[#ý-](.*)[#ý-]', 1, 1, NULL, 1) AS SUB_STRING
FROM cteData;
The parentheses around the .* between the delimiter groups makes the .* a sub-expression, and the final ,1 in the parameter list tells REGEXP_SUBSTR to give you back the value of sub-expression #1. Since there's only one sub-expression in the regular expression it gives you back the value of the .*, which is what you're looking for.
sqlfiddle here
How can we convert a string of any length into a comma separated string with comma after every n characters. I am using Oracle 10g and above. I tried with REGEXP_SUBSTR but couldn't get desired result.
e.g.: for below string comma after every 5 characters.
input:
aaaaabbbbbcccccdddddeeeeefffff
output:
aaaaa,bbbbb,ccccc,ddddd,eeeee,fffff,
or
aaaaa,bbbbb,ccccc,ddddd,eeeee,fffff
Thanks in advance.
This can be done with regexp_replace, like so:
WITH sample_data AS (SELECT 'aaaaabbbbbcccccdddddeeeeefffff' str FROM dual UNION ALL
SELECT 'aaaa' str FROM dual UNION ALL
SELECT 'aaaaabb' str FROM dual)
SELECT str,
regexp_replace(str, '(.{5})', '\1,')
FROM sample_data;
STR REGEXP_REPLACE(STR,'(.{5})','\
------------------------------ --------------------------------------------------------------------------------
aaaaabbbbbcccccdddddeeeeefffff aaaaa,bbbbb,ccccc,ddddd,eeeee,fffff,
aaaa aaaa
aaaaabb aaaaa,bb
The regexp_replace simply looks for any 5 characters (.{5}), and then replaces them with the same 5 characters plus a comma. The brackets around the .{5} turn it into a labelled subexpression - \1, since it's the first set of brackets - which we can then use to represent our 5 characters in the replacement section.
You would then need to trim the extra comma off the resultant string, if necessary.
SELECT RTRIM ( REGEXP_REPLACE('aaaaabbbbbcccccdddddeeeeefffff', '(.{5})' ,'\1,') ,',') replaced
FROM DUAL;
This worked for me:
WITH strlen AS
(
SELECT 'aaaaabbbbbcccccdddddeeeeefffffggggg' AS input,
LENGTH('aaaaabbbbbcccccdddddeeeeefffffggggg') AS LEN,
5 AS part
FROM dual
)
,
pattern AS
(
SELECT regexp_substr(strlen.input, '[[:alnum:]]{5}', 1, LEVEL)
||',' AS line
FROM strlen,
dual
CONNECT BY LEVEL <= strlen.len / strlen.part
)
SELECT rtrim(listagg(line, '') WITHIN GROUP (
ORDER BY 1), ',') AS big_bang$
FROM pattern ;
I have a string something like this 'SERO02~~~NA_#ERO5'. I need to sub string it using delimiter ~~~. So can get SERO02 and NA_#ERO5 as result.
I create an regex experession like this:
select regexp_substr('SERO02~~~NA_#ERO5' ,'[^~~~]+',1,2) from dual;
It worked fine and returns : NA_#ERO5
But if I change the string to ERO02~NA_#ERO5 the result is still same.
But I expect the expression to return nothing since delimiter ~~~ is not found in that string. Can someone help me out to create correct expression?
[^~~~] matches a single character that is not one of the characters following the caret in the square brackets. Since all those characters are identical then [^~~~] is the same as [^~].
You can match it using:
SELECT REGEXP_SUBSTR(
'SERO02~~~NA_#ERO5',
'~~~(.*?)(~~~|$)',
1,
1,
NULL,
1
)
FROM DUAL;
Which will match ~~~ then store zero-or-more characters in a capture group (the round brackets () indicates a capture group) until it finds either ~~~ or the end-of-string. It will then return the first capture group.
You can do it without regular expressions, with a bit of logics:
with test(text) as ( select 'SERO02~~~NA_#ERO5' from dual)
select case
when instr(text, '~~~') != 0 then
substr(text, instr(text, '~~~') + 3)
else
null
end
from test
This will give the part of the string after '~~~', if it exists, null otherwise.
You can edit the ELSE part to get what you need when the input string does not contain '~~~'.
Even using regexp,to match the string '~~~', you need to write it exactly, without []; the [] is used to list a set of characters, so [aaaaa] is exactly the same than [a],while [abc] means 'a' OR 'b' OR 'c'.
With regexp, even if not necessary, one way could be the following:
substr(regexp_substr(text, '~~~.*'), 4)
In case you want all elements. Handles NULL elements too:
SQL> with tbl(str) as (
select 'SERO02~~~NA_#ERO5' from dual
)
select regexp_substr(str, '(.*?)(~~~|$)', 1, level, null, 1) element
from tbl
connect by level <= regexp_count(str, '~~~') + 1;
ELEMENT
-----------------
SERO02
NA_#ERO5
SQL>
How to replace multiple whole characters, except those in combinations...?
The below code replaces multiple characters, but it also disturbing those in combinations.
SELECT regexp_replace('a,ca,va,ea,r,y,q,b,g','(a|y|q|g)','X') RESULT FROM dual;
Current output:
RESULT
--------------------
X,cX,vX,eX,r,X,X,b,X
Expected output:
RESULT
------------------------
'X,ca,va,ea,r,X,X,b,X
I just want to replace only separate whole characters('a','y','q','g'), but not the 1 in combinations('ca','va','ea')...
Because you are delimiting with a comma ',' you can combine that like ',a,'
and this will replace only single a's.
you can try follows:
with t as
(
select 'a,ca,va,ea,r,y,q,b,g' str
from dual
)
select substr(sys_connect_by_path(regexp_replace(regexp_substr(str, '[^,]+', 1, level), '^(a|y|q|g)$', 'X'), ','), 2) as str
from t
where connect_by_isleaf = 1
connect by level <= length(regexp_replace(str, '[^,]*')) + 1;
Sadly oracle doesn´t support lookahead and lookbehind. But this is a solution i came up with.
SELECT regexp_replace
(regexp_replace
('a,ca,va,ea,r,y,q,b,g',
'^[ayqg](,)|(,)[ayqg](,)|(,)[ayqg]$',
'\2\4X\1\3'),'(,)[ayqg](,)','\1X\2')
RESULT FROM dual;
I had to use the regexp twice sadly, since it doesn´t find two similar values following after each other and replacing it. ..,a,y,.. is getting replaced as ..,X,y,... So the second call replaces the missing [ayqg] with the exact values. In the first inner regexp call replaces the first and last values.
Maybe this could be simplified into one expression, but i am not that conform with the regex from oracle.
As a explanation i am grouping the commata and basicly replace every ,[ayqg], with ,X, by backreferencing the commata
You would look for word boundaries, which is \b, and which is unfortunately not supported by Oracle's regexp_replace.
So let's look for a non-word character \W or the beginning ^ or ending $ of the text.
select
regexp_replace('a,ca,va,ea,r,y,q,b,g','(^|$|\W)(a|y|q|g)(^|$|\W)','\1X\3') as result
from dual;
In order to not remove the non-word characters, we must have them in the replace string: \1 for the expression in the first parenteses, \3 for the ones in the third. Thus we only change the expression in the second parentheses, which is a, y, q or g, with X.
Unfortunately above gives
X,ca,va,ea,r,X,q,b,X
The q was not replaced, because we recognize ',y,' thus being positioned a 'g,' whereas we'd need to be positioned at ',g,' to recognize g as a word, too.
So we need to replace in iterations (i.e. recursively):
with results(txt, num) as
(
select 'a,ca,va,ea,r,y,q,b,g' as txt, 0 as num from dual
union all
select regexp_replace(txt, '(^|$|\W)(a|y|q|g)(^|$|\W)','\1X\3'), num + 1 as num
from results
where txt <> regexp_replace(txt, '(^|$|\W)(a|y|q|g)(^|$|\W)','\1X\3')
)
select max(txt) keep (dense_rank last order by num) as result
from results;
EDIT: Kevin Esche is right; of course one has to do it only twice. Hence you can also do:
select
regexp_replace(txt, search_str, replace_str) as result
from
(
select
regexp_replace(txt, search_str, replace_str) as txt, search_str, replace_str
from
(
select
'a,ca,va,ea,r,y,q,y,q,b,g' as txt,
'(^|$|\W)(a|y|q|g)(^|$|\W)' as search_str,
'\1X\3' as replace_str
from dual
)
);
with replaced_values as (
SELECT case when length(val)=1 then regexp_replace(val,'(a|y|q|g)','X') else val end new_val, lvl
from (
SELECT regexp_substr('a,ca,va,ea,r,y,q,b,g','[^,]+', 1, LEVEL) val, level lvl FROM dual
connect by regexp_substr('a,ca,va,ea,r,y,q,b,g','[^,]+',1, LEVEL) is not null
) all_values
)
select lISTAGG(new_val, ',') WITHIN GROUP (ORDER BY lvl) RESULT
from replaced_values
This statement pivots data into rows and replaces only lines wich contains one character.
Data are then unpivoted in one rows
This sql works also with empty entries like 'a,,,b,c' and more complex regular expressions:
with t as
(select ',a,,ca,va,ea,bbb,ba,r,y,q,b,g,,,' as str,
',' as delimiter,
'(a|y|q|g|ea|[b]*)' as regexp_expr,
'X' as replace_expr
from dual)
(select substr (sys_connect_by_path(regexp_replace(substr(str,
decode(level - 1, 0, 0, instr(str, ',', 1, level - 1)) + 1,
decode(instr(str, ',', 1, level),
0,
length(str),
instr(str, ',', 1, level) - 1) -
decode(level - 1, 0, 0, instr(str, ',', 1, level - 1))),
'^' || regexp_expr || '$',
replace_expr), ','), 2)
from t
where connect_by_isleaf = 1
connect by level <= length(regexp_replace(str, '[^'|| delimiter||']')) + 1)
Result
,X,,ca,va,X,X,ba,r,X,X,X,X,,,
Don't Know much Oracle, but I would have thought something like this could work. Assuming the delimiter is always a comma.
SELECT
regexp_replace(regexp_replace(regexp_replace(regexp_replace(regexp_replace('a,ca,va,ea,r,y,q,b,g','(,a,|,y,|,q,|,g,)',',X,') ,'(,a,|,y,|,q,|,g,)',',X,'), '(^a,|^y,|^q,|^g,)','X,'), '(,a$|,y$|,q$|,g$)',',X'), '(^a$|^y$|^q$|^g$)','X')
RESULT FROM test;
The first two parts replaces a single character in commas in the middle, the third part gets those at the start of the string, the fourth is for the end of the string and the fifth is for when then string has just one character.
This answer might will be simplifiable by advanced Regexp use.
How i can replace words?
RS & OS ===> D, LS & IS ==== >
SECTION_ID Output required
1-LS-1991 1-P-1991
1-IS-1991 1-P-1991
1-RS-1991 1- D- 1991
1-OS-1991 1-D-1991
I have below string:
ThisSentence.ShouldBe.SplitAfterLastPeriod.Sentence
So I want to select Sentence since it is the string after the last period. How can I do this?
Just for completeness' sake, here's a solution using regular expressions (not very complicated IMHO :-) ):
select regexp_substr(
'ThisSentence.ShouldBe.SplitAfterLastPeriod.Sentence',
'[^.]+$')
from dual
The regex
uses a negated character class to match anything except for a dot [^.]
adds a quantifier + to match one or more of these
uses an anchor $ to restrict matches to the end of the string
You can probably do this with complicated regular expressions. I like the following method:
select substr(str, - instr(reverse(str), '.') + 1)
Nothing like testing to see that this doesn't work when the string is at the end. Something about - 0 = 0. Here is an improvement:
select (case when str like '%.' then ''
else substr(str, - instr(reverse(str), ';') + 1)
end)
EDIT:
Your example works, both when I run it on my local Oracle and in SQL Fiddle.
I am running this code:
select (case when str like '%.' then ''
else substr(str, - instr(reverse(str), '.') + 1)
end)
from (select 'ThisSentence.ShouldBe.SplitAfterLastPeriod.Sentence' as str from dual) t
And yet another way.
Not sure from a performance standpoint which would be best...
The difference here is that we use -1 to count backwards to find the last . when doing the instr.
With CTE as
(Select 'ThisSentence.ShouldBe.SplitAfterLastPeriod.Sentence' str, length('ThisSentence.ShouldBe.SplitAfterLastPeriod.Sentence') len from dual)
Select substr(str,instr(str,'.',-1)+1,len-instr(str,'.',-1)+1) from cte;
select
substr(
'ThisSentence.ShouldBe.SplitAfterLastPeriod.Sentence',
INSTR(
'ThisSentence.ShouldBe.SplitAfterLastPeriod.Sentence',
'.',
-1
)+1
)
from dual;
The INSTR function accepts a third parameter, the occurrence. It defaults to 1 (the first occurrence), but also accepts negative numbers (meaning counting from the last occurrence backwards).
select substr(str, instr(str, '.', -1) + 1)
from (
select 'ThisSentence.ShouldBe.SplitAfterLastPeriod.Sentence'
as str
from dual);
Sentence
how many dots in a string?
select length(str) - length(replace(str, '.', '') number_of_dots from ...
get substring after last dot:
select substr(str, instr(str, '.', 1, number_of_dots)+1) from ...