I want to return a array in a function as my function looks like below,
CREATE OR REPLACE FUNCTION TEST
RETURN t_array
IS
strings t_array;
BEGIN
--do something
RETURN strings;
END:
But it gives a error t_array must be declare. I want to know where to declare it and how can i declare it?
When you are using a custom type, you have to declare that type first. For example:
CREATE OR REPLACE TYPE string_array IS TABLE OF varchar2(50);
Which creates a new type named string_array that is a table of varchars.
For more information check the official oracle dokumentation here
Related
I want to leave some fields empty (i.e. Null) when I insert values into table. I don't see why would I want to have a DB full of empty strings in fields.
I use Delphi 10, FireDAC and local SQLite DB.
Edit: Provided code is just an example. In my application values are provided by user input and functions, any many of them are optional. If value is empty, I would like to keep it at Null or default value. Creating multiple variants of ExecSQL and nesting If statements isn't an option too - there are too many optional fields (18, to be exact).
Test table:
CREATE TABLE "Clients" (
"Name" TEXT,
"Notes" TEXT
);
This is how I tried it:
var someName,someNote: string;
begin
{...}
someName:='Vasya';
someNote:='';
FDConnection1.ExecSQL('INSERT OR REPLACE INTO Clients(Name,Notes) VALUES (:nameval,:notesval)',
[someName, IfThen(someNote.isEmpty, Null, somenote)]);
This raises an exception:
could not convert variant of type (Null) into type (OleStr)
I've tried to overload it and specify [ftString,ftString] and it didn't help.
Currently I have to do it like this and I hate this messy code:
FDConnection1.ExecSQL('INSERT OR REPLACE INTO Clients(Name,Notes) VALUES ('+
IfThen(someName.isEmpty,'NULL','"'+Sanitize(someName)+'"')+','+
IfThen(someNote.isEmpty,'NULL','"'+Sanitize(someNote)+'"')+');');
Any recommendations?
Edit2: Currently I see an option of creating new row with "INSERT OR REPLACE" and then use multiple UPDATEs in a row for each non-empty value. But this looks direly ineffective. Like this:
FDConnection1.ExecSQL('INSERT OR REPLACE INTO Clients(Name) VALUES (:nameval)',[SomeName]);
id := FDConnection1.ExecSQLScalar('SELECT FROM Clients VALUES id WHERE Name=:nameval',[SomeName]);
if not SomeString.isEmpty then
FDConnection1.ExecSQL('UPDATE Clients SET Notes=:noteval WHERE id=:idval)',[SomeNote,id]);
According to Embarcadero documentation ( here ):
To set the parameter value to Null, specify the parameter data type,
then call the Clear method:
with FDQuery1.ParamByName('name') do begin
DataType := ftString;
Clear;
end;
FDQuery1.ExecSQL;
So, you have to use FDQuery to insert Null values, I suppose. Something like this:
//Assign FDConnection1 to FDQuery1's Connection property
FDQuery1.SQL.Text := 'INSERT OR REPLACE INTO Clients(Name,Notes) VALUES (:nameval,:notesval)';
with FDQuery1.ParamByName('nameval') do
begin
DataType := ftString;
Value := someName;
end;
with FDQuery1.ParamByName('notesval') do
begin
DataType := ftString;
if someNote.IsEmpty then
Clear;
else
Value := someNote;
end;
if not FDConnection1.Connected then
FDConnection.Open;
FDQuery1.ExecSql;
It's not very good idea to execute query as String without parameters because this code is vulnerable to SQL injections.
Some sources tells that it's not enough and you should do something like this:
with FDQuery1.ParamByName('name') do begin
DataType := ftString;
AsString := '';
Clear;
end;
FDQuery1.ExecSQL;
but I can't confirm it. You can try it if main example won't work.
how to manage the result of a query that returns an integer "select count(*) from table"?
1) I've tried to bind the output of a SQL Execute Statement service to an integer variable and doesn't work. (type mistmatch)
2) i've tried to use types like 'SQLResult', SQLResultRow, SQLResultColumn as well but they dont work:
Caused by: com.lombardisoftware.core.TeamWorksException: Type ismatch the value "[Element: ]" must be and instance of type atructured IBM BPM Java Class found: org.jdom.Element
3) i've tried to bind the output to a XMLElement variable and i've got this value
< resultSet recordCount=\"1\" columnCount=\"1\">5< /columnn>< /record>< /resultSet>
so now... how can I access the recordCount attribute of this node?
anyway, I don't like so manipulate a variable of XMLType, when are the types SQLResult, SQLResultRow, SQLResultColumn used?
****** EDITED *******
even if i get a result as XMLElement.. i can't manipulate it.
methods like: tw.local.result[0].rows[0].column[0].getText() don't work (the intellisense as well)
the XMLElement as an attribute "recordCount" but i don't know how to get his value..
Anyway, the only workaround that i found is to change the query in order to return a normal set of records(not a scalar value)
select field from table instead of select count(field) from table
so i could to map the output value to a list of objects and than count its length...
ugly and dirty :-(
anyone know how manipulate the XMLElement in a script block?
Please try this.
Bind the output variable from sql execute statement as 'ANY' type.
variable name - result (ANY)
Query - select count(field) as COUNTVAL from table
tw.local.totalCount = tw.local.result[0].rows[0].indexedMap.COUNTVAL;
Use Return type as XMLElement then bind a XMLElement in output mapping.
For eg: If you are using tw.local.output as ouput mapping (of type XMLElement) then,
log.info("Count "+tw.local.output.xpath('/resultSet/record/column').item(0).getText());
This will print the count
If you want to get "recordCount" Attribute then use
tw.local.output.getAttribute("recordCount");
I have to use existing stored procedure which returns REF CURSOR. I need to insert that resultset into a temporary table.
Spec of procedure is:
TYPE cur IS REF CURSOR;
PROCEDURE get(p_one NUMBER ,p_two OUT cur);
How can I insert the resultset of this procedure into a table.
I just re-read the title of your question. Do you actually need to be able to select from a procedure?
If so, this can be achieved by using pipelined functions.
The process for this is:
Create an object type to represent the record-type you require.
Create a nested table type of the object.
Create a pipelined function which returns the nest table.
You can then select from this function.
This example should get you on your way:
create or replace type to_test as object (
val1 varchar2(32),
val2 varchar2(32)
);
create or replace type tt_test as table of to_test;
create or replace function demo_pipe return tt_test pipelined
is
v_test to_test;
begin
for rec in (select * from user_tables) loop
v_test := to_test(rec.table_name, rec.tablespace_name);
pipe row (v_test);
end loop;
end;
/
select * from table(demo_pipe);
I'm designing a function that is part of a larger package. The function is intended to take a District Code and return a collection of unique IDs for 10-15 stores that are assigned to that district. The function is intended to return a collection that can be queried like a table, i.e., using the TABLE function in a SQL statement.
I've created the following Types:
Schema Level type:
create or replace TYPE HDT_CORE_ORGIDS AS TABLE OF CHAR(20);
and a Type inside the Package
TYPE CORE_ORGIDS IS TABLE OF CHAR(20) INDEX BY BINARY_INTEGER;
Here's the function code:
FUNCTION FindDistrictOrgs(
ParamOrgCode VARCHAR2
)
RETURN HDT_CORE_ORGIDS
AS
ReturnOrgs HDT_CORE_ORGIDS := HDT_CORE_ORGIDS();
FDOTemp HDT_CORE_MAIN.CORE_ORGIDS;
i BINARY_INTEGER := 0;
CURSOR FDOCurr IS
SELECT org.id AS OrgID
FROM tp2.tpt_company org
WHERE LEVEL = 2
START WITH org.name = ParamOrgCode
CONNECT BY PRIOR org.id = org.parent_id;
BEGIN
OPEN FDOCurr;
LOOP
i := i +1;
FETCH FDOCurr INTO FDOTemp(i);
EXIT WHEN FDOCurr%NOTFOUND;
END LOOP;
IF FDOTemp.EXISTS(FDOTemp.FIRST) THEN
ReturnOrgs.EXTEND(FDOTemp.LAST);
FOR x IN FDOTemp.FIRST .. FDOTemp.LAST LOOP
ReturnOrgs(x) := FDOTemp(x).OrgID;
END LOOP;
END IF;
CLOSE FDOCurr;
RETURN ReturnOrgs;
END FindDistrictOrgs ;
I'm getting the PLS-00487:Invalid Reference to variable 'CHAR' at the line:
ReturnOrgs(x) := FDOTemp(x).OrgID;
I've double-checked at the value returned by the SQL (the org.id AS OrgID) is of the CHAR(20 BYTE) datatype.
So...what's causing the error?
Any help is appreciated! :)
OrgID is the alias you gave the column in your cursor, it has no meaning to the collection. Since both collections are of simple types you should just be doing:
ReturnOrgs(x) := FDOTemp(x);
The syntax you're using is implying FDOTemp is a collection of objects and you're trying to reference the OrgID attribute of an object; but since CHAR isn't an object type, this errors. The error message even makes some sense when viewed like that, though it's not terribly helpful if you don't already know what's wrong... and not entirely helpful when you do.
Incidentally, you could use a bulk collect to populate the collection without the cursor or loops, or the extra collection:
SELECT org.id
BULK COLLECT INTO ReturnOrgs
FROM tp2.tpt_company org
WHERE LEVEL = 2
START WITH org.name = ParamOrgCode
CONNECT BY PRIOR org.id = org.parent_id;
RETURN ReturnOrgs;
Can anybody let me know how can i debug a functions with table type as input parameter and this function returns a table type pipelined.
Please see below details.When i try to test the function it creates below anonymous block but when i click on debug button it gives error:
Anonymous block:
declare
-- Non-scalar parameters require additional processing
result t_bmk_q;
pit_srch_str t_parm;
begin
-- Call the function
result := f_bmk_srch(pit_srch_str => pit_srch_str,
piv_op => 'ALL');
end;
---f_bmk_q function returns table type t_bmk_q pipelined
defintions:
==============
t_bmk_q --->table type
t_bmk_q is TABLE OF r_bmk_q -->object of some attributes.
pit_srch_str ---> is parameter of type t_parm which is table type of r_parm
--plz see def of r_parm:
CREATE OR REPLACE TYPE r_parm AS OBJECT
(
p_abc varchar2(200)
,p_new_val varchar2(2000)
,CONSTRUCTOR FUNCTION r_parm
(
p_abc varchar2
,p_new_val varchar2
) RETURN SELF AS RESULT
);
Example:I have below sample values to test and debug:
r_parm('TAB1.VALUE','123321123')
Thanks
Rajesh
It seems you are using a PL/SQL Developer Test Window to run the tests. I recognise the comment Non-scalar parameters require additional processing.
PL/SQL Developer's Test Window doesn't handle pipelined functions too well.
You are best off removing the result variable and wrapping the function call in open :cursor for select * from table(...). Add a variable named cursor of type Cursor to the variables list below the window.
To populate the input table, you can simply 'call' the table type, passing each row in the table as a separate argument. For example,
t_parm(r_parm(...), r_parm(...), r_parm(...))
You can add as many rows to the table as you like this way.
Putting both of these changes together, we have something like the following:
declare
pit_srch_str t_parm;
begin
-- Create input table.
pit_srch_str := t_parm(
r_parm('TAB1.VALUE','123321123'),
r_parm('TAB2.VALUE','456597646')
);
-- Call the function
open :cursor for select * from table(f_bmk_srch(pit_srch_str => pit_srch_str,
piv_op => 'ALL'));
end;
/
Once you've run the function, you can get the results from the cursor variable (use the ... button at the far right).