I got a question how to get number of characters from a string ?
In this scenario we got string: 'pppeerskka' and the outcome should look like '3p2e1r1s2k1a'
and the outcome should look like '3p2e1r1s2k1a'
One option is to split the input string into rows, apply count function to it (so that you'd know how many different letters you have), and - finally - aggregate them back using listagg.
SQL> with test (col) as
2 (select 'pppeerskka' from dual),
3 temp as
4 (select substr(col, level, 1) letter,
5 count(*) cnt,
6 max(level) lvl
7 from test
8 connect by level <= length(col)
9 group by substr(col, level, 1)
10 )
11 select listagg(cnt || letter, '') within group (order by lvl) result
12 from temp;
RESULT
--------------------------------------------------------------------------------
3p2e1r1s2k1a
SQL>
Converting it into a function is a simple task:
SQL> create or replace function f_count (par_string in varchar2)
2 return varchar2
3 is
4 retval varchar2(30);
5 begin
6 with temp as
7 (select substr(par_string, level, 1) letter,
8 count(*) cnt,
9 max(level) lvl
10 from test
11 connect by level <= length(par_string)
12 group by substr(par_string, level, 1)
13 )
14 select listagg(cnt || letter, '') within group (order by lvl)
15 into retval
16 from temp;
17
18 return retval;
19 end;
20 /
Function created.
SQL> select f_count('pppeerskka') result from dual;
RESULT
--------------------------------------------------------------------------------
3p2e1r1s2k1a
SQL>
Related
I don't know what is best solution for my problem. I need function with one parameter which resault is
VAL
-----
1
2
3
In function I need put union all to get all value.
Select column_1 as VAL from my_table where id = P_FUNCTION_PARAMETER --return 1
union all
Select column_2 as VAL from my_table where id = P_FUNCTION_PARAMETER --return 2
union all
Select column_3 as VAL from my_table where id = P_FUNCTION_PARAMETER; --return 3
What is the best solution for this?
"Best" just depends. Return a ref cursor or a collection, whichever you prefer.
For example:
SQL> create or replace function f_test_rc
2 return sys_refcursor
3 is
4 rc sys_refcursor;
5 begin
6 open rc for
7 select 1 from dual union all
8 select 2 from dual union all
9 select 3 from dual;
10
11 return rc;
12 end;
13 /
Function created.
SQL> select f_test_rc from dual;
F_TEST_RC
--------------------
CURSOR STATEMENT : 1
CURSOR STATEMENT : 1
1
----------
1
2
3
SQL> create or replace function f_test_coll
2 return sys.odcinumberlist
3 as
4 l_coll sys.odcinumberlist;
5 begin
6 select * bulk collect into l_coll
7 from (select 1 from dual union all
8 select 2 from dual union all
9 select 3 from dual
10 );
11
12 return l_coll;
13 end;
14 /
Function created.
SQL> select * from table(f_test_coll);
COLUMN_VALUE
------------
1
2
3
SQL>
First let's set up a small table for testing:
create table my_table
( id number primary key
, column_1 number
, column_2 number
, column_3 number
);
insert into my_table
select 1008, 3, -8, 0.2 from dual union all
select 1002, 6, null, -1.2 from dual
;
commit;
The function can look like this. Note that I don't use union all - that will require reading the table three times, when only one time is enough.
create or replace function my_function (p_function_parameter number)
return sys.odcinumberlist
as
arr sys.odcinumberlist;
begin
select case ord when 1 then column_1
when 2 then column_2
when 3 then column_3 end
bulk collect into arr
from my_table cross join
(select level as ord from dual connect by level <= 3)
where id = p_function_parameter
order by ord
;
return arr;
end;
/
The function could be used, for example, like this: (in older versions you may need to wrap the function call within the table operator)
select * from my_function(1002);
COLUMN_VALUE
------------
6
-1.2
I am trying to split data from a column into rows but I am facing this issue here.
When i run this query it splits the data fine but it is also returning NULL as an extra row too.
Here is the value I am trying to split ,162662163,90133140,163268955,169223426,169222899,
WITH CTE AS(
SELECT
RTRIM(LTRIM(PG2.MULTILIST11, ','), ',') ACCESS_BY_GROUPS
FROM AGILE.ITEM I
INNER JOIN AGILE.PAGE_TWO PG2 ON PG2.ID = I.ID
WHERE
ITEM_NUMBER IN --('313-000074',
('313-000090')
)
SELECT DISTINCT
REGEXP_SUBSTR(ACCESS_BY_GROUPS, '[^,]+', 1, column_value) ACCESS_BY_GROUPS
FROM CTE
CROSS JOIN TABLE(CAST(MULTISET(SELECT LEVEL FROM DUAL
CONNECT BY LEVEL <= REGEXP_COUNT(CTE.ACCESS_BY_GROUPS, ',') + 10
) AS sys.odcinumberlist))
I don't want to get that null value. I cannot apply that check out of the query because it will affect other column values too so I want it to be handled somewhere in the function. I hope someone can help.
It is because of leading and trailing commas.
One option is to begin from position 2 (see line #3) and limit number of values (line #5)
SQL> with test (col) as
2 (select ',162662163,90133140,163268955,169223426,169222899,' from dual)
3 select regexp_substr(col, '[^,]+', 2, level) res
4 from test
5 connect by level <= regexp_count(col, ',') - 1
6 /
RES
-------------------------------------------------
162662163
90133140
163268955
169223426
169222899
SQL>
Another is to remove leading and trailing comma first, and then split the rest into rows:
SQL> with test (col) as
2 (select ',162662163,90133140,163268955,169223426,169222899,' from dual),
3 temp as
4 -- remove leading and trailing comma first
5 (select ltrim(rtrim(col, ','), ',') col
6 from test
7 )
8 select regexp_substr(col, '[^,]+', 1, level) res
9 from temp
10 connect by level <= regexp_count(col, ',') + 1;
RES
------------------------------------------------
162662163
90133140
163268955
169223426
169222899
SQL>
[EDIT - for more than a single row]
If there are more rows involved, code has to be changed. Note that there must be some kind of a unique identifier for every row (ID in my example).
SQL> with test (id, col) as
2 (select 1, ',162662163,90133140,163268955,169223426,169222899,' from dual union all
3 select 2, ',1452761,1452762,' from dual
4 )
5 select id,
6 regexp_substr(col, '[^,]+', 2, column_value) res
7 from test cross join table(cast(multiset(select level from dual
8 connect by level <= regexp_count(col, ',') - 1
9 ) as sys.odcinumberlist))
10 order by id, column_value
11 /
ID RES
---------- -------------------------------------------------
1 162662163
1 90133140
1 163268955
1 169223426
1 169222899
2 1452761
2 1452762
7 rows selected.
SQL>
For example i have the table called 'Table1'. and column called 'country'.
I want to count the value of word in string.below is my data for column 'country':
country:
"japan singapore japan chinese chinese chinese"
expected output: in above data we can see the japan appear two time, singapore once and chinese 3 times.i want to count value of word where japan is count as one, singapore as one and chinese as one. hence the ouput will be 3. please help me
ValueOfWord: 3
Firstly, it is a bad design to store multiple values in a single column as delimited string. You should consider normalizing the data as a permanent solution.
With the denormalized data, you could do it in a single SQL using REGEXP_SUBSTR:
SELECT COUNT(DISTINCT(regexp_substr(country, '[^ ]+', 1, LEVEL))) as "COUNT"
FROM table_name
CONNECT BY LEVEL <= regexp_count(country, ' ')+1
/
Demo:
SQL> WITH sample_data AS
2 ( SELECT 'japan singapore japan chinese chinese chinese' str FROM dual
3 )
4 -- end of sample_data mocking real table
5 SELECT COUNT(DISTINCT(regexp_substr(str, '[^ ]+', 1, LEVEL))) as "COUNT"
6 FROM sample_data
7 CONNECT BY LEVEL <= regexp_count(str, ' ')+1
8 /
COUNT
----------
3
See Split single comma delimited string into rows in Oracle to understand how the query works.
UPDATE
For multiple delimited string rows you need to take care of the number of rows formed by the CONNECT BY clause.
See Split comma delimited strings in a table in Oracle for more ways of doing the same task.
Setup
Let's say you have a table with 3 rows like this:
SQL> CREATE TABLE t(country VARCHAR2(200));
Table created.
SQL> INSERT INTO t VALUES('japan singapore japan chinese chinese chinese');
1 row created.
SQL> INSERT INTO t VALUES('singapore indian malaysia');
1 row created.
SQL> INSERT INTO t VALUES('french french french');
1 row created.
SQL> COMMIT;
Commit complete.
SQL> SELECT * FROM t;
COUNTRY
---------------------------------------------------------------------------
japan singapore japan chinese chinese chinese
singapore indian malaysia
french french french
Using REGEXP_SUBSTR and REGEXP_COUNT:
We expect the output as 6 since there are 6 unique strings.
SQL> SELECT COUNT(DISTINCT(regexp_substr(t.country, '[^ ]+', 1, lines.column_value))) count
2 FROM t,
3 TABLE (CAST (MULTISET
4 (SELECT LEVEL FROM dual
5 CONNECT BY LEVEL <= regexp_count(t.country, ' ')+1
6 ) AS sys.odciNumberList ) ) lines
7 ORDER BY lines.column_value
8 /
COUNT
----------
6
There are many other methods to achieve the desired output. Let's see how:
Using XMLTABLE
SQL> SELECT COUNT(DISTINCT(country)) COUNT
2 FROM
3 (SELECT trim(COLUMN_VALUE) country
4 FROM t,
5 xmltable(('"'
6 || REPLACE(country, ' ', '","')
7 || '"'))
8 )
9 /
COUNT
----------
6
Using MODEL clause
SQL> WITH
2 model_param AS
3 (
4 SELECT country AS orig_str ,
5 ' '
6 || country
7 || ' ' AS mod_str ,
8 1 AS start_pos ,
9 Length(country) AS end_pos ,
10 (LENGTH(country) -
11 LENGTH(REPLACE(country, ' '))) + 1 AS element_count ,
12 0 AS element_no ,
13 ROWNUM AS rn
14 FROM t )
15 SELECT COUNT(DISTINCT(Substr(mod_str, start_pos, end_pos-start_pos))) count
16 FROM (
17 SELECT *
18 FROM model_param
19 MODEL PARTITION BY (rn, orig_str, mod_str)
20 DIMENSION BY (element_no)
21 MEASURES (start_pos, end_pos, element_count)
22 RULES ITERATE (2000)
23 UNTIL (ITERATION_NUMBER+1 = element_count[0])
24 ( start_pos[ITERATION_NUMBER+1] =
25 instr(cv(mod_str), ' ', 1, cv(element_no)) + 1,
26 end_pos[ITERATION_NUMBER+1] =
27 instr(cv(mod_str), ' ', 1, cv(element_no) + 1) )
28 )
29 WHERE element_no != 0
30 ORDER BY mod_str , element_no
31 /
COUNT
----------
6
Did you store that kind of string in a single entry?
If not, try
SELECT COUNT(*)
FROM (SELECT DISTINCT T.country FROM Table1 T)
If yes, I would write an external program to parse the string and return the result you want.
Like using java.
Create a String set.
I would use JDBC to retrieve the record, and use split to split strings in tokens using ' 'delimiter. For every token, if it is not in the set, add it to the set.
When parse finishes, get the length of the set, which is the value you want.
Break the string based on the space delimiter
SELECT COUNT(DISTINCT regexp_substr(col, '[^ ]+', 1, LEVEL))
FROM T
CONNECT BY LEVEL <= regexp_count(col, ' ')+1
For counting DISTINCT words
SELECT col,
COUNT(DISTINCT regexp_substr(col, '[^ ]+', 1, LEVEL))
FROM T
CONNECT BY LEVEL <= regexp_count(col, ' ')+1
GROUP BY col
FIDDLE
I have values like below
1,a,b,c
2,d,e
3,f,g
Expected output
1 a
1 b
1 c
2 d
2 e
Can you please help me?
It is very much close to implementing the logic to Split comma delimited strings in a table. The only tricky thing is that you have the row number along with the string itself.
You could use ROWNUM as pseudo column, and then filter out those rows where the leading substr of the string is repeating with the ROWNUM.
For example,
Setup
SQL> CREATE TABLE t(text VARCHAR2(4000));
Table created.
SQL>
SQL> INSERT INTO t SELECT '1,a,b,c' text FROM dual;
1 row created.
SQL> INSERT INTO t SELECT '2,d,e' text FROM dual;
1 row created.
SQL> INSERT INTO t SELECT '3,f,g' text FROM dual;
1 row created.
SQL> COMMIT;
Commit complete.
SQL>
SQL> SELECT * FROM t;
TEXT
----------
1,a,b,c
2,d,e
3,f,g
SQL>
Solution:
SQL> WITH DATA AS(
2 SELECT ROWNUM rn, text FROM t
3 )
4 SELECT *
5 FROM
6 (SELECT rn,
7 trim(regexp_substr(t.text, '[^,]+', 1, lines.COLUMN_VALUE)) text
8 FROM DATA t,
9 TABLE (CAST (MULTISET
10 (SELECT LEVEL FROM dual CONNECT BY LEVEL <= regexp_count(t.text, ',')+1
11 ) AS sys.odciNumberList ) ) lines
12 )
13 WHERE TO_CHAR(rn) <> text
14 ORDER BY rn
15 /
RN TEXT
---------- ----------
1 a
1 b
1 c
2 d
2 e
3 f
3 g
7 rows selected.
SQL>
when I select the table from Oracle, I want to handle one col'val :
eg:
'ab,cd,ef' to 'ef->cd->ab';
'AB,BC' to 'BC->AB';
'ACNN,BBCCAC' to 'BBCCAC->ACNN';
'BBBDC，DCCX，FFF' to 'FFF->DCCX->BBBDC'
We have two tasks. The first is to tokenize the original strings. This is quite easy with regular expressions (although there are more performant approaches if you are dealing with large volumes). The second task is to re-assemble the tokens in reverse order; we can use the 11gR2 LISTAGG() function for this:
with tokens as (
select distinct col1, regexp_substr(col1, '[^,]+', 1, level) as tkn, level as rn
from t23
connect by level <= regexp_count (col1, '[,]') +1
)
select col1
, listagg(tkn, '->')
within group (order by rn desc) as rev_col1
from tokens
group by col1
/
Here is a SQL Fiddle.
You can do it with a mix of string split and string aggregation.
Using:
REGEXP_SUBSTR : To split the comma delimited string into rows
LISTAGG : To aggregate the values
You can have a look at this article to understand how string split works http://lalitkumarb.wordpress.com/2015/03/04/split-comma-delimited-strings-in-a-table-using-oracle-sql/
SQL> WITH DATA AS(
2 SELECT 1 ID, 'ab,cd,ef' text FROM dual UNION ALL
3 SELECT 2 ID, 'AB,BC' text FROM dual UNION ALL
4 SELECT 3 ID, 'ACNN,BBCCAC' text FROM dual
5 )
6 SELECT ID,
7 listagg(text, ',') WITHIN GROUP (
8 ORDER BY rn DESC) reversed_indices
9 FROM
10 (SELECT t.id,
11 rownum rn,
12 trim(regexp_substr(t.text, '[^,]+', 1, lines.COLUMN_VALUE)) text
13 FROM data t,
14 TABLE (CAST (MULTISET
15 (SELECT LEVEL FROM dual CONNECT BY LEVEL <= regexp_count(t.text, ',')+1
16 ) AS sys.odciNumberList ) ) lines
17 ORDER BY ID
18 )
19 GROUP BY ID
20 /
ID REVERSED_INDICES
---------- ------------------------------
1 ef,cd,ab
2 BC,AB
3 BBCCAC,ACNN
SQL>
Let's say your table looks like:
SQL> SELECT * FROM t;
ID TEXT
---------- ------------------------------
1 ab,cd,ef
2 AB,BC
3 ACNN,BBCCAC
4 word1,word2,word3
5 1,2,3
SQL>
Using the above query:
SQL> SELECT ID,
2 listagg(text, '-->') WITHIN GROUP (
3 ORDER BY rn DESC) reversed_indices
4 FROM
5 (SELECT t.id,
6 rownum rn,
7 trim(regexp_substr(t.text, '[^,]+', 1, lines.COLUMN_VALUE)) text
8 FROM t,
9 TABLE (CAST (MULTISET
10 (SELECT LEVEL FROM dual CONNECT BY LEVEL <= regexp_count(t.text, ',')+1
11 ) AS sys.odciNumberList ) ) lines
12 ORDER BY ID
13 )
14 GROUP BY ID
15 /
ID REVERSED_INDICES
---------- ------------------------------
1 ef-->cd-->ab
2 BC-->AB
3 BBCCAC-->ACNN
4 word3-->word2-->word1
5 3-->2-->1
SQL>