how to get a unqiue result sets in PL/SQL cursor? - plsql

I want use this procedure to display the username and moblephone number,the result sets is this when I use select :
declare enter image description here
when the procedure runs,I get this :
enter image description here
error ORA-01722: invalid number
ORA-06512: at "ABPROD.SHAREPOOL", line 24.
when I use unique or distinct in the cursor,nothing display.
the code source :
create or replace procedure sharepool (assignment in varchar2,myorgname in varchar2) is
rightid T_CLM_AP30_RIGHT.RIGHT_ID%type;
orgid t_clm_ap30_org.org_id%type;
begin
select t.right_id into rightid from T_CLM_AP30_RIGHT t where t.rightdesc=trim(assignment);
dbms_output.put_line(rightid||trim(myorgname)||assignment);
select t.org_id into orgid from t_clm_ap30_org t where t.orgname=trim(myorgname);
dbms_output.put_line(orgid);
declare
cursor namelist is select distinct a.username,a.mobile from t_clm_ap30_user a, T_CLM_AP30_RIGHT_AUTH t where a.user_id=t.user_id and t.right_id=rightid and t.poolorgrange=orgid ;
begin
for c in namelist
loop
dbms_output.put_line(c.username||' '||c.mobile);
end loop;
end;
end sharepool;

INVALID_NUMBER errors indicate a failed casting of a string to a number. That means one of your join conditions is comparing a string column with a number column, and you have values in the string column which cannot be cast to a number.
ORA-06512: at "ABPROD.SHAREPOOL", line 24
Line 24 doesn't align with the code you've posted, presumably lost in translation from your actual source. Also you haven't posted table descriptions so we cannot tell which columns to look at.
So here is a guess.
One (or more) of these joins has an implicit numeric conversion:
where a.user_id = t.user_id
and t.right_id = rightid
and t.poolorgrange = orgid
That is, either t_clm_ap30_user.user_id is numeric and T_CLM_AP30_RIGHT_AUTH.user_id is not, or vice versa. Or T_CLM_AP30_RIGHT_AUTH.right_id is numeric and T_CLM_AP30_RIGHT.right_id is not, or vice versa. Or T_CLM_AP30_RIGHT_AUTH.poolorgrange is numeric and t_clm_ap30_org.org_id is not, or vice versa.
Only you can figure this out, because only you can see your schema. Once you have identified the join where you have a string column being compared to a numeric column you need to query that column to find the data which cannot be converted to a number.
Let's say that T_CLM_AP30_RIGHT_AUTH.poolorgrange is the rogue string. You can see which are the troublesome rows with this query:
select * from T_CLM_AP30_RIGHT_AUTH
where translate (poolorgrange, 'x1234567890', 'x') is not null;
The translate() function strips out digits. So anything which is left can't be converted to a number.
"T_CLM_AP30_RIGHT_AUTH.poolorgrange is varchar2,and t_clm_ap30_org.org_id is numeric ."
You can avoid the error by explicitly casting the t_clm_ap30_org.org_id to a string:
select distinct a.username, a.mobile
from t_clm_ap30_user a,
T_CLM_AP30_RIGHT_AUTH t
where a.user_id = t.user_id
and t.right_id = rightid
and t.poolorgrange = to_char(orgid) ;
Obviously you're not going to get matches on those alphanumeric values but the query will run.

Related

expression is of wrong type for function

CREATE OR REPLACE FUNCTION k_w_b_salary(k IN NUMBER, b IN BOOLEAN)
RETURN EMP.ENAME%TYPE IS
names name_table;
BEGIN
IF (b = true) THEN
SELECT ENAME BULK COLLECT INTO names
FROM
(SELECT *
FROM EMP
ORDER BY SAL ASC)
WHERE ROWNUM <= k;
RETURN names;
ELSIF (b = false) THEN
SELECT ENAME BULK COLLECT INTO names
FROM
(SELECT *
FROM EMP
ORDER BY SAL DESC)
WHERE ROWNUM <= k;
RETURN names;
END IF;
END;
And I get this error:
12/9 PL/SQL: Statement ignored
12/16 PLS-00382: expression is of wrong type
20/9 PL/SQL: Statement ignored
20/16 PLS-00382: expression is of wrong type
I have this function that tries to find the best/worst paid employees. But i get the above error.
I think it's something to do with the ROWNUM but I am not sure.
I think the lines the error points out are not the lines with the error.
I had this function writen differently and the lines in the error where pointing to the ROWNUM <= k lines.
I have tried putting a fixed number there (<= 3) for example and I got the same error.
I have no idea what else to try, i can't really understand why this is not working.
It's not obvious to me why this is not working. I think it should work fine but obviously it dousen't.
The code for the table i use is :
CREATE OR REPLACE TYPE name_table IS TABLE OF VARCHAR2(10);
Any help is appreciated!
In the function declaration, you said
RETURN EMP.ENAME%TYPE
I assume the data type of column ENAME in table EMP is some sort of string (VARCHAR2(40) or similar) - right?
In the declarations section, you declare a variable names of data type name_table. You didn't show us the definition of the name_table type (that must be given outside the function, not in it); we can probably assume it is a nested table of some sort. Right? [EDIT - I take that back; you did show us your definition of name_table, at the end of your question.]
In the end, your function returns names. Which is of type name_table. But you said the function returns something else: EMP.ENAME%TYPE. In particular, you said the function returns a scalar data type, but you are returning a collection.
This will not work even if the collection has a single element. A table with a single "record" is not the same data type as the "record" itself - even if an actual table has a single "record" in it.
(And, much more so, when the table has three records in it!)
Rather: It seems that you want a table function: one that returns a table of things. If so, then declare the function that way. Perhaps you want the function to
RETURN NAME_TABLE
(at the top, in the function declaration)

Reading 'bad' Sqlite data with FireDAC

I just learned that the Sqlite Manager plug-in for Firefox will go away in November, so I've been trying to recreate its functionality in Delphi: open Sqlite databases, enter SQL queries. I'm running Tokyo.
My problem comes with Sqlite fields defined as 'date.' While Sqlite allows specification of data types, it allows pretty much anything to be put in any field. FireDAC handles bad entries in integer or float fields ('bar' becomes 0, '32foo' becomes 32), but it hiccups on fields described as 'date.'
For example, I have a table:
CREATE TABLE "someTable" ("id" INTEGER, "s" text(10), "d" date)
With this data:
INSERT INTO "someTable" VALUES ("1","good date","2017-09-09");
INSERT INTO "someTable" VALUES ("2","bad date","2017-09-0b");
INSERT INTO "someTable" VALUES ("3","empty d","");
INSERT INTO "someTable" VALUES ("4","null date",null);
Opening FDQuery q1 with SQL = "select * from someTable", a bad date (such as the second record) raises an EConvertError ("Invalid argument to date encode"). I tried to get around it by adding a maprule (q1 is a FDQuery):
with q1.FormatOptions do begin
OwnMapRules := True;
with MapRules.Add do begin
SourceDataType := dtDate;
TargetDataType := dtAnsiString;
sizemin := 10;
sizemax := 256;
PrecMin := -1;
PrecMax := -1;
ScaleMin := -1;
ScaleMax := -1;
end;
end;
However, that raises an EFDException:
[FireDAC][DatS]-32. Variable length column overflow. Value length - [10], column maximum length - [0].
What am I missing?
Why do I get "Invalid argument to date encode" exception when fetching tuple containing invalid DATE type field value?
The problem is that FireDAC expects DATE data type values represented as a string in fixed format YYYY-MM-DD where all members must be integers. Internally used FDStr2Int function doesn't use any detection of invalid input and works directly with ASCII ordinary values shifted by the 0 char, so input like e.g. GHIJ-KL-MN results in year 25676, month 298, day 320 after parsing. And just these misinterpreted values are then passed to the EncodeDate function, which fails for such values with the exception you've mentioned:
Invalid argument to date encode
The above happens inside the GetData method (ParseDate nested function). One possible way for fixing this issue on FireDAC's side could be using safer function TryEncodeDate instead of a direct encoding attempt. Similar problem is with TIME data type string value.

use sqlite check to validate whether date with proper format is entered in the column

I have created a table as below:
CREATE TABLE case_status(data_entry_timestamp DATETIME DEFAULT (datetime('now','localtime')) NOT NULL,
case_number TEXT PRIMARY KEY NOT NULL,
case_name TEXT DEFAULT MISSING,
death_reportdate DATE CONSTRAINT death_reportdate_chk CHECK (death_reportdate==strftime('%Y-%m-%d',death_reportdate)),
);
The column death_reportdate need to have a date with pre-defined format (e.g. 2000-12-31). I created the table, inserted some rows of data, and then try to modified data in death_reportdate, the check rule seems to be bypassed when I enter some random string to it.
What have I done wrong?
You had an extra comma at the end. Correct code:
CREATE TABLE case_status(data_entry_timestamp DATETIME DEFAULT (datetime('now','localtime')) NOT NULL,
case_number TEXT PRIMARY KEY NOT NULL,
case_name TEXT DEFAULT MISSING,
death_reportdate DATE CONSTRAINT death_reportdate_chk CHECK (death_reportdate==strftime('%Y-%m-%d',death_reportdate))
)
it is an old Topic but i had the the same Problem. if the strftime method Fails to Format the string( a bad Input) it retuns null, so you have to check is not null in the end
Here is another solution which works like a charm:
`date` DATE CHECK(date IS strftime('%Y-%m-%d', date))
This also works with the time:
`time` TIME CHECK(time IS strftime('%H:%M:%S', time))
Use this to define your column. I think that is a more elegant solution than checking for null value.
First, two small notes.
I'm using the TEXT type since SQLite does not have "real types." It has 5 column "affinities", INTEGER, TEXT, BLOB, REAL, and NUMERIC. If you say DATE then it uses NUMERIC which can behave a little weirdly in my opinion. I find it best to explicitly use one of the 5 affinities.
I'm using date(...) instead of strftime('%Y-%m-%d', ...) because they are the same thing.
Let's break down why the original question did not work.
DROP TABLE IF EXISTS TEMP.example;
CREATE TEMPORARY TABLE example (
deathdate TEXT CHECK (deathdate == date(deathdate))
);
INSERT INTO TEMP.example (deathdate) VALUES ('2020-01-01');
INSERT INTO TEMP.example (deathdate) VALUES ('a');
INSERT INTO TEMP.example (deathdate) VALUES (NULL);
SELECT * FROM TEMP.example;
Running this lets all three values get into the database. Why? Let's check the documentation for CHECK constraints.
If the result is zero (integer value 0 or real value 0.0), then a constraint violation has occurred. If the CHECK expression evaluates to NULL, or any other non-zero value, it is not a constraint violation.
If you run SELECT 'a' == date('a'); you'll see it is NULL. Why? Check SELECT date('a'); and you'll see it is also NULL. Huh, maybe the documentation for == can help?
Note that there are two variations of the equals and not equals operators. Equals can be either = or ==. [...]
The IS and IS NOT operators work like = and != except when one or both of the operands are NULL. In this case, if both operands are NULL, then the IS operator evaluates to 1 (true) and the IS NOT operator evaluates to 0 (false). If one operand is NULL and the other is not, then the IS operator evaluates to 0 (false) and the IS NOT operator is 1 (true). It is not possible for an IS or IS NOT expression to evaluate to NULL.
We need to use IS, not ==, and trying that we see that 'a' no longer gets in.
DROP TABLE IF EXISTS TEMP.example;
CREATE TEMPORARY TABLE example (
deathdate TEXT CHECK (deathdate IS date(deathdate))
);
INSERT INTO TEMP.example (deathdate) VALUES ('2020-01-01');
INSERT INTO TEMP.example (deathdate) VALUES ('a');
INSERT INTO TEMP.example (deathdate) VALUES (NULL);
SELECT * FROM TEMP.example;
If you don't want NULL to get in, simple change it to deathdate TEXT NOT NULL CHECK (deathdate IS date(deathdate))

PL/SQL - LAST_VALUE return more than one row?

I am doing an school assigment where I need to get the last value of "code" so I can then insert next row with this code incremented. I tried to pull it out this way.
DECLARE
v_last_code f_shifts.code%TYPE;
BEGIN
SELECT LAST_VALUE(code) OVER (ORDER BY code)
INTO v_last_code
FROM f_shifts;
DBMS_OUTPUT.PUT_LINE('Last value is: ' || v_last_code);
END;
However I get ORA-01422: exact fetch returns more than one requested number of rows
and I have no idea why and how can a last_value be more than one row
Thanks !
You can use a nested table like this.
DECLARE
v_last_code f_shifts.code%TYPE;
TYPE t_tbl IS TABLE OF f_shifts.code%TYPE;
-- Above line creates the nested table type of the required type.
v_tbl T_TBL;
-- Above line creates the nested table variable.
BEGIN
SELECT code
BULK COLLECT INTO v_tbl -- collects all the values, ordered, into the nested table
FROM f_shifts
ORDER BY code;
v_last_code = v_tbl(v_tbl.LAST); -- gets the last values and puts into the variable
DBMS_OUTPUT.PUT_LINE('Last value is: ' || v_last_code);
END;
/

SQLite SELECT statement where column equals zero

I'm preety new to SQLite.
I have a preety basic question.. Why can't I select rows where specific column equals zero?
The is_unwanted column is type TINYINT (which I see in SQLite basically means INTEGER)
So, I have only one record in the database (for testing).
When I try
SELECT is_unwanted FROM 'urls'
I get a result of "0" (zero), which is fine because that column contains the actual number 0.
I tried =>
SELECT * FROM 'urls' WHERE is_unwanted = 0
And got NO result, but
SELECT * FROM 'urls' WHERE is_unwanted <> 0
gives me result.
What am I doing wrong??
Try running
select '{' || is_unwanted || '}' from urls
to see if the value in the database is really a string containing spaces.
SQLite is a dynamically typed database; when you specify TINYINT is is a hint (SQLite uses the term "affinity") for the column. You can use
select is_unwanted, typeof(is_unwanted) from urls
to see the values with their types.
You could try:
SELECT * FROM urls WHERE coalesce(is_unwanted,'') = ''

Resources