I have a table with this values:
ID VALUE
-----------------------
23559 200
23562 -1 & {14376}#-1
and I want to do to a select that if I cannot convert to number set NULL.
I generally use translate for this because it is such an odd corner case:
SELECT
CASE
WHEN NOT TRIM(TRANSLATE(COLUMN_NAME, '1234567890', ' ')) IS NULL THEN NULL
ELSE COLUMN_NAME
END AS "NUMERIC_COLUMN"
FROM
TABLE_NAME;
If necessary, that can be turned into a procedure, but I'm not sure that there would be terribly much benefit performance-wise.
You can create a function that tries to convert the string to a number and catches the exception. Something like
CREATE OR REPLACE FUNCTION my_to_number( p_str IN VARCHAR2 )
RETURN NUMBER
IS
l_num NUMBER;
BEGIN
BEGIN
l_num := to_number( p_str );
EXCEPTION
WHEN others THEN
l_num := null;
END;
RETURN l_num;
END;
Then you can
SELECT id, my_to_number( value )
FROM your_table
You could also use REGEXP_LIKE:
SELECT id
, CASE WHEN regexp_like(value,'^[0-9]+$') THEN TO_NUMBER(value)
ELSE NULL
END value
FROM your_table;
For example:
SQL> WITH q AS (
2 SELECT 1 ID, '200' col FROM dual
3 UNION
4 SELECT 2, '-1 & {14376}#-1' FROM dual
5 )
6 SELECT id, CASE WHEN regexp_like(col,'^[0-9]+$') THEN TO_NUMBER(col) ELSE NULL END TEST FROM q;
ID TEST
---------- ----------
1 200
2
With Oracle 12.2 this can be done a bit easier using the on conversion error option:
select id, cast(value as number default null on conversion error) as value
from the_table;
Optionally you can also specify a format mask, similar to the to_number() function.
I assume this would be faster than using a PL/SQL function, not sure about the performance compared to a case with a regex. But it is definitely a lot shorter.
CREATE OR REPLACE FUNCTION asnumber(p_val IN VARCHAR2) RETURN NUMBER IS
l_val NUMBER;
BEGIN
l_val := TO_NUMBER(p_val);
RETURN l_val;
EXCEPTION WHEN VALUE_ERROR THEN
RETURN null;
END;
Related
function that take two parameters, the first to be a string and the second is the order (Asc or Desc) and the returned output to be ordering the first string as per the second parameter.
IN : dgtak
OUT: adgkt
Tried this but doesn't seem to work
CREATE OR REPLACE FUNCTION order_string(my_string IN VARCHAR2)
RETURN VARCHAR2 IS
ret_string VARCHAR2(4000);
BEGIN
SELECT LISTAGG(regexp_substr(my_string, '\w', 1, level), '') WITHIN
GROUP(
ORDER BY 1)
INTO ret_string
FROM dual
CONNECT BY regexp_substr(my_string, '\w', 1, level) IS NOT NULL;
RETURN ret_string;
END;
select order_string('dgtak') as RESULT from dual;
Here's one option:
SQL> create or replace function order_string (par_string in varchar2, par_order in varchar2)
2 return varchar2
3 is
4 retval varchar2(100);
5 begin
6 with temp (val) as
7 -- split PAR_STRING to rows
8 (select substr(par_string, level, 1)
9 from dual
10 connect by level <= length(par_string)
11 )
12 -- aggregate characters back in ascending or descending order
13 select case when par_order = 'Asc' then listagg(val, '') within group (order by val asc)
14 when par_order = 'Desc' then listagg(val, '') within group (order by val desc)
15 else null
16 end
17 into retval
18 from temp;
19
20 return retval;
21 end;
22 /
Function created.
Testing:
SQL> select order_string('dfag', 'Asc') result_asc,
2 order_string('dfag', 'Desc') result_desc
3 from dual;
RESULT_ASC RESULT_DESC
-------------------- --------------------
adfg gfda
SQL>
Just for fun, here's a procedural version. It has more lines of code than the SQL version but in my tests it's slightly faster.
create or replace function order_string
( p_string varchar2
, p_reverse varchar2 default 'N' )
return varchar2
as
pragma udf;
type letter_tt is table of number index by varchar2(1);
letters letter_tt := letter_tt();
letter varchar2(1);
sorted_string long;
string_length integer := length(p_string);
begin
-- Store all characters of p_string as indices of array:
for i in 1..string_length loop
letter := substr(p_string,i,1);
if letters.exists(letter) then
letters(letter) := letters(letter) +1;
else
letters(letter) := 1;
end if;
end loop;
-- Loop through array appending each array index to sorted_string
for i in indices of letters loop
for r in 1..letters(i) loop
sorted_string := sorted_string || i;
end loop;
end loop;
if p_reverse = 'Y' then
select reverse(sorted_string) into sorted_string from dual;
end if;
return sorted_string;
end order_string;
I've used the 21c indices of loop iterator, but you can write a conventional loop in earlier versions. You might also use two alternative loops for ascending and descending order in place of my hack.
I have a table called 'config' and when I query it in following manner:
SELECT value FROM config WHERE property = 'SPECIAL_STORE_ID'
its response will be: 59216;131205;76707;167206 //... (1)
I want to tokenize the above values using semicolon as the delimiter and then use them in a user-defined Function's IF statement to compare, something like this:
IF in_store_id exists in (<delimited response from (1) above>)//...(2)
THEN do some stuff
where in_store_id is the parameter passed-in to the function
Is this possible to do as one-liner in (2) above ?
I'm on Oracle 12c
One-liner? I don't think so, but - if you're satisfied with something like this, fine.
SQL> select * From config;
VALUE PROPERTY
-------------- ----------------
7369;7499;7521 SPECIAL_STORE_ID
SQL> declare
2 in_store_id varchar2(20) := 7369;
3 l_exists number;
4 begin
5 select instr(value, ';' || in_store_id || ';')
6 into l_exists
7 from config
8 where property = 'SPECIAL_STORE_ID';
9
10 if l_exists > 0 then
11 dbms_output.put_line('that STORE_ID exists in the value');
12 else
13 dbms_output.put_line('that STORE_ID does not exist in the value');
14 end if;
15 end;
16 /
that STORE_ID exists in the value
PL/SQL procedure successfully completed.
SQL>
If the delimited response is a collection then you can use member of to check if the collection contains the ID or not like
create or replace procedure test_procedure2(p_property in varchar2, p_id in varchar2) is
type test_t is table of varchar2(20);
l_ids test_t;
begin
select regexp_substr(value, '[^;]+', 1, level) bulk collect into l_ids
from (select value from config where property = p_property)
connect by level <= regexp_count(value, ';')+1;
if(p_id member of (l_ids)) then
dbms_output.put_line('Do stuff for '||p_property||' '||p_id);
end if;
end;
/
or do it without the collection with intermediate select like
create or replace procedure test_procedure1(p_property in varchar2, p_id in varchar2) is
l_flag number(3);
begin
select count(1) into l_flag from dual where p_id in (
select regexp_substr(value, '[^;]+', 1, level)
from (select value from config where property = p_property)
connect by level <= regexp_count(value, ';')+1
);
if(l_flag > 0) then
dbms_output.put_line('Do stuff for '||p_property||' '||p_id);
end if;
end;
/
See fiddle
I have a table "test_calculate" this has a column "CONN_BY" having values
column can have more than 2 number to multiply and this table may contain millions of rows , I need to get the result of the calculation from "CONN_BY" to "MVP".
I have used xmlquery for the calculation and dynamic query but these are quite slow. Is there another way which is much faster .Please suggest.
You can try the dynamic query.
Create a function which returns the calculated value and use it in your insert or select queries.
CREATE OR REPLACE FUNCTION UFN_CALCULATE (CLM_VALUE VARCHAR2)
RETURN NUMBER IS
RES_VAL NUMBER;
BEGIN
EXECUTE IMMEDIATE 'select '||CLM_VALUE||' FROM DUAL' INTO RES_VAL;
RETURN RES_VAL;
END;
You can use that function like below.
SELECT UFN_CALCULATE('.0876543 * .09876') FROM DUAL;
SELECT UFN_CALCULATE(CONN_BY) FROM YOUR_TABLE;
One option is using select ... connect by level <= regexp_count(conn_by,'[^*]+')... query for the implicit cursor within a PL/SQL code block
SQL> set serveroutput on
SQL> declare
mvp owa.nc_arr; -- numeric array to initialize each multiplication to 1 for each id value
begin
dbms_output.put_line('ID MVP');
dbms_output.put_line('--------');
for c in
(
select id,
to_number( regexp_substr(conn_by,'[^*]+',1,level) ) as nr,
level as lvl , max( level ) over ( partition by id ) as mx_lvl
from test_calculate
connect by level <= regexp_count(conn_by,'[^*]+')
and prior sys_guid() is not null
and prior conn_by = conn_by
order by id, lvl
)
loop
if c.lvl = 1 then mvp(c.id) := 1; end if;
mvp(c.id) := c.nr * mvp(c.id);
if c.lvl = c.mx_lvl then
dbms_output.put_line(c.id||' '||mvp(c.id));
end if;
end loop;
end;
/
where test_calculate is assumed to have an identity column(id)
Demo
I want to use my result of function e.g. 'S500,S600,S700,S800' in a subquery in another script like:
where dept_no in (my result of function)
So I want to convert my string result to be like this ('S500','S600','S700','S800').
I tried to do this with dynamic SQL but I can't get it to work.
Hope below snipet suffice your requirement.
Approach 1 -> More effective
--Create a table type of VARCHAR
CREATE OR REPLACE type string_table
IS
TABLE OF VARCHAR2(100);
--Function to return tabl type
CREATE OR REPLACE
FUNCTION string_manipulate
RETURN string_table
AS
str_tab string_table;
BEGIN
SELECT 's00'||level bulk collect INTO str_tab FROM dual CONNECT BY level < 10;
RETURN str_tab;
end;
--Use function in the query
SELECT distinct 1
FROM
(SELECT 's001' dn FROM dual
UNION ALL
SELECT 's002' dn FROM dual
UNION ALL
SELECT 's003' dn FROM dual
UNION ALL
SELECT 's004' dn FROM dual
UNION ALL
SELECT 's005' dn FROM dual
UNION ALL
SELECT 's006' dn FROM dual
UNION ALL
SELECT 's007' dn FROM dual
UNION ALL
SELECT 's008' dn FROM dual
UNION ALL
SELECT 's009' dn FROM dual
)a
WHERE a.dn IN
(SELECT * FROM TABLE(string_manipulate)
);
--Approach 2
--Function to get output as mentioned.
CREATE OR REPLACE
FUNCTION string_manipulate
RETURN VARCHAR2
AS
BEGIN
RETURN 'S2009,S2020,S2021';
END;
-- Use function value in a query
SELECT 1
FROM dual
WHERE '''S2009'',''S2020'',''S2021''' = (''''
||REPLACE(string_manipulate,',',''',''')
||'''');
You need an iterator and text splitting by comma sign:
select empno,ename,sal,deptno
from emp
where empno in (
select to_number(
rtrim(
substr(emps,
instr(emps,',',1,iter.pos)+1,
instr(emps,',',1,iter.pos+1) -
instr(emps,',',1,iter.pos)),',')) emps
from (select ','||'7654,7698,7782,7788'||',' emps from t1) csv,
(select rownum pos from emp) iter
where iter.pos <= ((length(csv.emps) -
length(replace(csv.emps,',')))/length(','))-1
)
But better rewrite your function to return cursor.
you can use collection:
SELECT *
FROM YOUR_TABLE
WHERE DEPT_NO IN (SELECT *
FROM TABLE (SPLIT ('S500,S600,S700,S800')))--splits text with comma, for other chars use split(text, split_char)
With usage of MEMBER OF
SELECT *
FROM YOUR_TABLE
WHERE DEPT_NO MEMBER OF SPLIT ('S500,S600,S700,S800')--splits text with comma, for other chars use split(text, split_char)
the split fuction is:
CREATE OR REPLACE TYPE SPLIT_TBL AS TABLE OF VARCHAR2 (32767);
CREATE OR REPLACE FUNCTION SPLIT (P_LIST VARCHAR2, P_DEL VARCHAR2 := ',')
RETURN SPLIT_TBL
PIPELINED
IS
L_IDX PLS_INTEGER;
L_LIST VARCHAR2 (32767) := P_LIST;
BEGIN
LOOP
L_IDX := INSTR (L_LIST, P_DEL);
IF L_IDX > 0
THEN
PIPE ROW (SUBSTR (L_LIST, 1, L_IDX - 1));
L_LIST := SUBSTR (L_LIST, L_IDX + LENGTH (P_DEL));
ELSE
PIPE ROW (L_LIST);
EXIT;
END IF;
END LOOP;
RETURN;
END SPLIT;
FUNCTION GET_TS_EACH_DAY_DEPARTMENT (P_SER_NO VARCHAR2,
P_TS_DATE DATE
)
RETURN STRING_TABLE
IS
V_DEPT_NO VARCHAR2 (4000);
V_DEPT VARCHAR2(4000);
V_TABLE STRING_TABLE:=STRING_TABLE();
J NUMBER:=1;
BEGIN
for i in (select distinct ts_day dayy from WEB_TS_USER_LOCATIONS_V ) loop
V_TABLE.EXTEND;
V_TABLE(J):= WEB_TS_PKG.GET_TS_DAY_DEPARTMENT (P_SER_NO ,P_TS_DATE , i.dayy );
J:=J+1;
end loop;
RETURN V_TABLE;
END GET_TS_EACH_DAY_DEPARTMENT;
Without using PL/SQL, is it possible to do data masking in SELECT statement?
For example:
(AS-IS) SELECT 'this is a string' from DUAL;
this is a string
(TO-BE) SELECT 'this is a string' from DUAL;
xxxx xx x xxxxxx
REGEXP_REPLACE can do this:
SELECT REGEXP_REPLACE('this is a string', '\w', 'x') FROM DUAL;
This replaces all non-whitespace characters with an x. To replace letters only, try this:
SELECT REGEXP_REPLACE('this is a string', '[A-Za-z]', 'x') FROM DUAL;
You can create user defined function as below and call that function in your query to mask the data.
create or replace function scrubbing(word in varchar2)
return varchar2
as
each_var char(2);
final_val varchar2(100);
complete_data varchar2(4000);
each_word varchar2(1000);
cursor val is select substr(replace(word,' ','#'),-level,1) from dual connect by level<=length(word);
begin
open val;
--final_val:= '';
loop
fetch val into each_var;
exit when val%NOTFOUND;
--dbms_output.put_line(each_var);
final_val := trim(final_val)||trim(each_var);
--dbms_output.put_line(final_val);
select regexp_substr(final_val,'[A-Za-z]+') into each_word from dual;
select replace(translate(final_val,each_word,dbms_random.string('L',length(word))),'#',' ') into complete_data from dual;
end loop;
return complete_data;
end;