I would like to execute a query that will only show all the string before dash in the particular field.
For example:
Original data: AB-123
After query: AB
You can use substr:
SQL> WITH DATA AS (SELECT 'AB-123' txt FROM dual)
2 SELECT substr(txt, 1, instr(txt, '-') - 1)
3 FROM DATA;
SUBSTR(TXT,1,INSTR(TXT,'-')-1)
------------------------------
AB
or regexp_substr (10g+):
SQL> WITH DATA AS (SELECT 'AB-123' txt FROM dual)
2 SELECT regexp_substr(txt, '^[^-]*')
3 FROM DATA;
REGEXP_SUBSTR(TXT,'^[^-]*')
---------------------------
AB
You can use regexp_replace.
For example
WITH DATA AS (
SELECT 'AB-123' as text FROM dual
UNION ALL
SELECT 'ABC123' as text FROM dual
)
SELECT
regexp_replace(d.text, '-.*$', '') as result
FROM DATA d;
will lead to
WITH DATA AS (
2 SELECT 'AB-123' as text FROM dual
3 UNION ALL
4 SELECT 'ABC123' as text FROM dual
5 )
6 SELECT
7 regexp_replace(d.text, '-.*$', '') as result
8 FROM DATA d;
RESULT
------------------------------------------------------
AB
ABC123
I found this simple
SELECT distinct
regexp_replace(d.pyid, '-.*$', '') as result
FROM schema.table d;
pyID column contains ABC-123, DEF-3454
SQL Result:
ABC
DEF
Related
I have a string in Oracle DB in the format 'a|b|c' , '|' being the separator between characters. Want to write an SQL query to transform it into a string in the format 'a,b&c'. First occurrence of '|' to ',' 2nd occurrence to '&'.
If suppose the string comes in the format 'a|b' then output should be 'a&b'.
I'm using multiple regex_replace queries to achieve this right now.
select REGEXP_REPLACE ('a|b|c', '[|]', ',', 1, 1)
from dual
Is there any other solution using one single query?
Nested replaces (see line #6):
SQL> with test (col) as
2 (select 'a|b|c' from dual union all
3 select 'a|b' from dual
4 )
5 select col,
6 regexp_replace(regexp_replace(col, '\|', ',', 1, 1), '\|', '&', 2, 1) result
7 from test;
COL RESULT
----- --------------------
a|b|c a,b&c
a|b a,b
SQL>
If you want to replace the last | in the list with & and all the preceding |s with , then you can use:
SELECT value,
REPLACE(
REGEXP_REPLACE( value, '\|([^|]*)$', '&\1' ),
'|',
','
) AS replaced
FROM table_name;
Which, for the sample data:
CREATE TABLE table_name (value) AS
SELECT 'A|B|C|D' FROM DUAL UNION ALL
SELECT 'A|B|C' FROM DUAL UNION ALL
SELECT 'A|B' FROM DUAL UNION ALL
SELECT 'A' FROM DUAL;
Outputs:
VALUE
REPLACED
A|B|C|D
A,B,C&D
A|B|C
A,B&C
A|B
A&B
A
A
fiddle
Table_A:
AA
BB
CC
DD
Output:
AA,BB,CC,DD
Any one help this
LISTAGG seems to be one choice:
Sample data:
SQL> with table_a(col) as
2 (select 'AA' from dual union all
3 select 'BB' from dual union all
4 select 'CC' from dual union all
5 select 'DD' from dual
6 )
Query:
7 select listagg(col, ',') within group (order by col) result
8 from table_a;
RESULT
------------------------------
AA,BB,CC,DD
SQL>
informatica only solution -
SRT : sort the data by col1.
EXP : create 3 ports. make sure their length be >3000 char
in_col1 = col1
v_col1 = col1|v_col1
p_col1= col1
o_col1= v_col1
AGG: create 2 ports
in_col1 <-- link o_col1 from previous step2
o_col1 = MAX(in_col1) -- link this to final target
Normally, there should be a key column by which you want to aggregate the strings. if you have any, you need to change step2 in v_col1.
enter image description hereWhat happens if if ship 3 and 4 are null, but ship2 is not null, that should be city state
Here is sample data in the picture.
I prefer the oldfashioned SUBSTR + INSTR combination which, if compared to Gordon's and Barbaros' suggestions, seems to be somewhat better as their queries return strings that don't even contain a comma, while the OP says
extract city from 1 letter until 1 comma
Here's a comparison:
SQL> with tab (addr) as
2 (
3 select 'RALEIGH, NC 27604-3229' from dual union all
4 select 'SUITE A' from dual union all
5 select 'COEUR D ALENE, ID 83815-8652' from dual union all
6 select '*O/S CITY LIMITS*' from dual
7 )
8 select addr,
9 substr(addr, 1, instr(addr, ',') - 1) littlefoot,
10 --
11 regexp_substr(addr, '[^,]+', 1, 1) gordon,
12 regexp_substr(addr,'[^,]+') barbaros
13 from tab;
ADDR LITTLEFOOT GORDON BARBAROS
---------------------------- --------------- -------------------- --------------------
RALEIGH, NC 27604-3229 RALEIGH RALEIGH RALEIGH
SUITE A SUITE A SUITE A
COEUR D ALENE, ID 83815-8652 COEUR D ALENE COEUR D ALENE COEUR D ALENE
*O/S CITY LIMITS* *O/S CITY LIMITS* *O/S CITY LIMITS*
SQL>
If you want the part before the first comma, you can use regexp_substr():
select regexp_substr(addr, '[^,]+', 1, 1)
Just use regexp_substr with [^,]+ pattern as below
select regexp_substr(address,'[^,]+') as city
from tab;
SQL Fiddle Demo 1
Or alternatively by creating an auxilary table :
with tab as
(
select 'RALEIGH, NC 27604-3229' as str from dual union all
select 'SALINAS, CA 93901' from dual union all
select 'DEPEW, NY 14043-2603' from dual
)
select regexp_substr(str,'[^,]+') as city
from tab;
SQL Fiddle Demo 2
If you don't want to use regexp, you can just use:
select substr(city,1,(instr(city,',')-1))
from mytable;
I have data like this in a table
NAME PRICE
A 2
B 3
C 5
D 9
E 5
I want to display all the values in one row; for instance:
A,2|B,3|C,5|D,9|E,5|
How would I go about making a query that will give me a string like this in Oracle? I don't need it to be programmed into something; I just want a way to get that line to appear in the results so I can copy it over and paste it in a word document.
My Oracle version is 10.2.0.5.
-- Oracle 10g --
SELECT deptno, WM_CONCAT(ename) AS employees
FROM scott.emp
GROUP BY deptno;
Output:
10 CLARK,MILLER,KING
20 SMITH,FORD,ADAMS,SCOTT,JONES
30 ALLEN,JAMES,TURNER,BLAKE,MARTIN,WARD
I know this is a little late but try this:
SELECT LISTAGG(CONCAT(CONCAT(NAME,','),PRICE),'|') WITHIN GROUP (ORDER BY NAME) AS CONCATDATA
FROM your_table
Usually when I need something like that quickly and I want to stay on SQL without using PL/SQL, I use something similar to the hack below:
select sys_connect_by_path(col, ', ') as concat
from
(
select 'E' as col, 1 as seq from dual
union
select 'F', 2 from dual
union
select 'G', 3 from dual
)
where seq = 3
start with seq = 1
connect by prior seq+1 = seq
It's a hierarchical query which uses the "sys_connect_by_path" special function, which is designed to get the "path" from a parent to a child.
What we are doing is simulating that the record with seq=1 is the parent of the record with seq=2 and so fourth, and then getting the full path of the last child (in this case, record with seq = 3), which will effectively be a concatenation of all the "col" columns
Adapted to your case:
select sys_connect_by_path(to_clob(col), '|') as concat
from
(
select name || ',' || price as col, rownum as seq, max(rownum) over (partition by 1) as max_seq
from
(
/* Simulating your table */
select 'A' as name, 2 as price from dual
union
select 'B' as name, 3 as price from dual
union
select 'C' as name, 5 as price from dual
union
select 'D' as name, 9 as price from dual
union
select 'E' as name, 5 as price from dual
)
)
where seq = max_seq
start with seq = 1
connect by prior seq+1 = seq
Result is: |A,2|B,3|C,5|D,9|E,5
As you're in Oracle 10g you can't use the excellent listagg(). However, there are numerous other string aggregation techniques.
There's no particular need for all the complicated stuff. Assuming the following table
create table a ( NAME varchar2(1), PRICE number);
insert all
into a values ('A', 2)
into a values ('B', 3)
into a values ('C', 5)
into a values ('D', 9)
into a values ('E', 5)
select * from dual
The unsupported function wm_concat should be sufficient:
select replace(replace(wm_concat (name || '#' || price), ',', '|'), '#', ',')
from a;
REPLACE(REPLACE(WM_CONCAT(NAME||'#'||PRICE),',','|'),'#',',')
--------------------------------------------------------------------------------
A,2|B,3|C,5|D,9|E,5
But, you could also alter Tom Kyte's stragg, also in the above link, to do it without the replace functions.
Here is another approach, using model clause:
-- sample of data from your question
with t1(NAME1, PRICE) as(
select 'A', 2 from dual union all
select 'B', 3 from dual union all
select 'C', 5 from dual union all
select 'D', 9 from dual union all
select 'E', 5 from dual
) -- the query
select Res
from (select name1
, price
, rn
, res
from t1
model
dimension by (row_number() over(order by name1) rn)
measures (name1, price, cast(null as varchar2(101)) as res)
(res[rn] order by rn desc = name1[cv()] || ',' || price[cv()] || '|' || res[cv() + 1])
)
where rn = 1
Result:
RES
----------------------
A,2|B,3|C,5|D,9|E,5|
SQLFiddle Example
Something like the following, which is grossly inefficient and untested.
create function foo returning varchar2 as
(
declare bar varchar2(8000) --arbitrary number
CURSOR cur IS
SELECT name,price
from my_table
LOOP
FETCH cur INTO r;
EXIT WHEN cur%NOTFOUND;
bar:= r.name|| ',' ||r.price || '|'
END LOOP;
dbms_output.put_line(bar);
return bar
)
Managed to get till here using xmlagg: using oracle 11G from sql fiddle.
Data Table:
COL1 COL2 COL3
1 0 0
1 1 1
2 0 0
3 0 0
3 1 0
SELECT
RTRIM(REPLACE(REPLACE(
XMLAgg(XMLElement("x", col1,',', col2, col3)
ORDER BY col1), '<x>'), '</x>', '|')) AS COLS
FROM ab
;
Results:
COLS
1,00| 3,00| 2,00| 1,11| 3,10|
* SQLFIDDLE DEMO
Reference to read on XMLAGG
For these strings
RSLR_AIRL19_ID3454_T20030913091226
RSLR_AIRL19_ID3122454_T20030913091226
RSLR_AIRL19_ID34_T20030913091226
How to get the number after ID ?
Or how to get the content between two characters but not include them ?
I use this '/\_ID([^_]+)/' got matches like Array ( [0] => _ID3454 [1] => 3454 )
Is this the right way?
To extract a number after an ID, you could write a similar query.
SQL> with t1 as(
2 select 'RSLR_AIRL19_ID3454_T20030913091226' as col from dual union all
3 select 'RSLR_AIRL19_ID3122454_T20030913091226' from dual union all
4 select 'RSLR_AIRL19_ID34_T20030913091226' from dual
5 )
6 select regexp_substr(col, '^([[:alnum:]]+_){2}ID([[:digit:]]+)_([[:alnum:]]+){1}$', 1, 1, 'i', 2) as ID
7 from t1
8 ;
ID
-------------
3454
3122454
34
Or, if you want to extract digits from a first occurrence of the pattern without verifying if an entire string matches a specific format:
SQL> with t1 as(
2 select 'RSLR_AI_RL19_ID3454_T20030913091226' as col from dual union all
3 select 'RSLR_AIRL19_ID3122454_T20030913091226' from dual union all
4 select 'RSLR_AIRL19_ID34_T20030913091226' from dual
5 )
6 select regexp_substr(col, 'ID([[:digit:]]+)', 1, 1, 'i', 1) as ID
7 from t1
8 ;
ID
--------------
3454
3122454
34
With pcre & perl engines :
ID\K\w+
NOTE
\K "restart" the match.
See http://www.phpfreaks.com/blog/pcre-regex-spotlight-k (php use pcre)