I am new to Postgres, in ms sql server we can write stored procedure to perform logic and return a select statement. Likewise i wrote following pqsql function with return type table. But its showing some incorrect syntax error, but if I replace the return type with integer and comment out the select statement it works fine.
following is the function with return type table
CREATE OR REPLACE FUNCTION candidate_save
(
p_name varchar(50),
p_dob date,
p_course_level_code integer,
p_email varchar(50),
p_mob character(10),
p_sslc_regno varchar(10),
p_sslc_year_of_passing character(4),
p_sslc_board_of_examination integer,
p_password character(128),
p_ip_address varchar(15)
)
RETURNS TABLE
(
regno char(10),
name varchar(50),
dob date,
course_level_code integer,
email varchar(50),
mob character(10),
sslc_regno varchar(10),
sslc_year_of_passing character(4),
sslc_boe integer
)
AS
$BODY$
DECLARE
max_row_count integer;
random_no integer;
v_regno character(10);
BEGIN
SELECT COUNT(*) INTO max_row_count FROM candidates WHERE course_level_code = p_course_level_code;
max_row_count := max_row_count + 1;
v_regno := cast(p_course_level_code as character) || cast( trunc(random() * 89999 + 10000) as integer) || to_char(max_row_count, 'FM0000');
INSERT INTO candidates
(regno, name, dob, course_level_code, email, mob, sslc_regno, sslc_year_of_passing, sslc_board_of_examination, password, created_on, ip_address)
VALUES
(v_regno, p_name, p_dob, p_course_level_code, p_email, p_mob, p_sslc_regno, p_sslc_year_of_passing, p_sslc_board_of_examination, p_password, now(), p_ip_address);
RETURN QUERY
SELECT regno, name, dob, course_level_name, email, mob, sslc_regno, sslc_year_of_passing, c.board as sslc_boe
FROM candidates a
INNER JOIN course_levels b on a.course_level_code = b.course_level_code
INNER JOIN sslc_board_of_examinations c ON a.sslc_board_of_examination = boe_code
WHERE regno = v_regno;
END;
$BODY$
LANGUAGE plpgsql;
when creating above function following error is showing
ERROR: syntax error at or near "$1"
LINE 1: INSERT INTO candidates ( $1 , $2 , $3 , $4 , $5 , $6 , ...
^
QUERY: INSERT INTO candidates ( $1 , $2 , $3 , $4 , $5 , $6 , $7 , $8 , sslc_board_of_examination, password, created_on, ip_address) VALUES ( $9 , $10 , $11 , $12 , $13 , $14 , $15 , $16 , $17 , $18 , now(), $19 )
CONTEXT: SQL statement in PL/PgSQL function "candidate_save" near line 30
********** Error **********
ERROR: syntax error at or near "$1"
SQL state: 42601
Context: SQL statement in PL/PgSQL function "candidate_save" near line 30
but if I replace the table with integer it works fine.
CREATE OR REPLACE FUNCTION candidate_save
(
p_name varchar(50),
p_dob date,
p_course_level_code integer,
p_email varchar(50),
p_mob character(10),
p_sslc_regno varchar(10),
p_sslc_year_of_passing character(4),
p_sslc_board_of_examination integer,
p_password character(128),
p_ip_address varchar(15)
)
RETURNS integer
AS
$BODY$
DECLARE
max_row_count integer;
random_no integer;
v_regno character(10);
BEGIN
SELECT COUNT(*) INTO max_row_count FROM candidates WHERE course_level_code = p_course_level_code;
max_row_count := max_row_count + 1;
v_regno := cast(p_course_level_code as character) || cast( trunc(random() * 89999 + 10000) as integer) || to_char(max_row_count, 'FM0000');
INSERT INTO candidates
(regno, name, dob, course_level_code, email, mob, sslc_regno, sslc_year_of_passing, sslc_board_of_examination, password, created_on, ip_address)
VALUES
(v_regno, p_name, p_dob, p_course_level_code, p_email, p_mob, p_sslc_regno, p_sslc_year_of_passing, p_sslc_board_of_examination, p_password, now(), p_ip_address);
RETURN 1;
END;
$BODY$
LANGUAGE plpgsql;
if I am doing something wrong pl. let me know.
I am using Postgres 9.4
INSERT INTO candidates
(regno
^------- The error
ERROR: syntax error at or near "$1"
LINE 1: INSERT INTO candidates ( $1
The issue is that regno is a plpgsql variable, because it's a column name in the output table.
RETURNS TABLE
(
regno char(10),
Each entry in RETURNS TABLE is registered as a variable so you can assign to it and use RETURN NEXT.
When parsed and processed, plpgsql variables in statements are replaced with positional parameters like $1, $2, etc. Which is why the error is what it is, and also why some of the later entries in the insert column-name-list don't get replaced. They don't clash with a parameter name.
Use different variable names.
I think 9.5 detects parameter name clashes and reports a clearer error.
finally it started working after I changed the variable names in the return table.
Thanks to all
I am importing data from a table which has raw feeds in Varchar, I need to import a column in varchar into a string column. I tried using the <column_name>::integer as well as to_number(<column_name>,'9999999') but I am getting errors, as there are a few empty fields, I need to retrieve them as empty or null into the new table.
Wild guess: If your value is an empty string, you can use NULLIF to replace it for a NULL:
SELECT
NULLIF(your_value, '')::int
You can even go one further and restrict on this coalesced field such as, for example:-
SELECT CAST(coalesce(<column>, '0') AS integer) as new_field
from <table>
where CAST(coalesce(<column>, '0') AS integer) >= 10;
If you need to treat empty columns as NULLs, try this:
SELECT CAST(nullif(<column>, '') AS integer);
On the other hand, if you do have NULL values that you need to avoid, try:
SELECT CAST(coalesce(<column>, '0') AS integer);
I do agree, error message would help a lot.
The only way I succeed to not having an error because of NULL, or special characters or empty string is by doing this:
SELECT REGEXP_REPLACE(COALESCE(<column>::character varying, '0'), '[^0-9]*' ,'0')::integer FROM table
I'm not able to comment (too little reputation? I'm pretty new) on Lukas' post.
On my PG setup to_number(NULL) does not work, so my solution would be:
SELECT CASE WHEN column = NULL THEN NULL ELSE column :: Integer END
FROM table
If the value contains non-numeric characters, you can convert the value to an integer as follows:
SELECT CASE WHEN <column>~E'^\\d+$' THEN CAST (<column> AS INTEGER) ELSE 0 END FROM table;
The CASE operator checks the < column>, if it matches the integer pattern, it converts the rate into an integer, otherwise it returns 0
Common issue
Naively type casting any string into an integer like so
SELECT ''::integer
Often results to the famous error:
Query failed: ERROR: invalid input syntax for integer: ""
Problem
PostgreSQL has no pre-defined function for safely type casting any string into an integer.
Solution
Create a user-defined function inspired by PHP's intval() function.
CREATE FUNCTION intval(character varying) RETURNS integer AS $$
SELECT
CASE
WHEN length(btrim(regexp_replace($1, '[^0-9]', '','g')))>0 THEN btrim(regexp_replace($1, '[^0-9]', '','g'))::integer
ELSE 0
END AS intval;
$$
LANGUAGE SQL
IMMUTABLE
RETURNS NULL ON NULL INPUT;
Usage
/* Example 1 */
SELECT intval('9000');
-- output: 9000
/* Example 2 */
SELECT intval('9gag');
-- output: 9
/* Example 3 */
SELECT intval('the quick brown fox jumps over the lazy dog');
-- output: 0
you can use this query
SUM(NULLIF(conversion_units, '')::numeric)
And if your column has decimal points
select NULLIF('105.0', '')::decimal
This works for me:
select (left(regexp_replace(coalesce('<column_name>', '0') || '', '[^0-9]', '', 'g'), 8) || '0')::integer
For easy view:
select (
left(
regexp_replace(
-- if null then '0', and convert to string for regexp
coalesce('<column_name>', '0') || '',
'[^0-9]',
'',
'g'
), -- remove everything except numbers
8 -- ensure ::integer doesn't overload
) || '0' -- ensure not empty string gets to ::integer
)::integer
The perfect solution for me is to use nullif and regexp_replace
SELECT NULLIF(REGEXP_REPLACE('98123162t3712t37', '[^0-9]', '', 'g'), '')::bigint;
Above solution consider the following edge cases.
String and Number: only the regexp_replace function perfectly converts into integers.
SELECT NULLIF(REGEXP_REPLACE('string and 12345', '[^0-9]', '', 'g'), '')::bigint;
Only string: regexp_replace converts non-string characters to empty strings; which can't cast directly to integer so use nullif to convert to null
SELECT NULLIF(REGEXP_REPLACE('only string', '[^0-9]', '', 'g'), '')::bigint;
Integer range: Converting a string into integer may cause out of range for type integer error. So use bigint instead
SELECT NULLIF(REGEXP_REPLACE('98123162t3712t37', '[^0-9]', '', 'g'), '')::bigint;
I'm looking for a built-in function/extended function in T-SQL for string manipulation similar to the String.Format method in .NET.
If you are using SQL Server 2012 and above, you can use FORMATMESSAGE. eg.
DECLARE #s NVARCHAR(50) = 'World';
DECLARE #d INT = 123;
SELECT FORMATMESSAGE('Hello %s, %d', #s, #d)
-- RETURNS 'Hello World, 123'
More examples from MSDN: FORMATMESSAGE
SELECT FORMATMESSAGE('Signed int %i, %d %i, %d, %+i, %+d, %+i, %+d', 5, -5, 50, -50, -11, -11, 11, 11);
SELECT FORMATMESSAGE('Signed int with leading zero %020i', 5);
SELECT FORMATMESSAGE('Signed int with leading zero 0 %020i', -55);
SELECT FORMATMESSAGE('Unsigned int %u, %u', 50, -50);
SELECT FORMATMESSAGE('Unsigned octal %o, %o', 50, -50);
SELECT FORMATMESSAGE('Unsigned hexadecimal %x, %X, %X, %X, %x', 11, 11, -11, 50, -50);
SELECT FORMATMESSAGE('Unsigned octal with prefix: %#o, %#o', 50, -50);
SELECT FORMATMESSAGE('Unsigned hexadecimal with prefix: %#x, %#X, %#X, %X, %x', 11, 11, -11, 50, -50);
SELECT FORMATMESSAGE('Hello %s!', 'TEST');
SELECT FORMATMESSAGE('Hello %20s!', 'TEST');
SELECT FORMATMESSAGE('Hello %-20s!', 'TEST');
SELECT FORMATMESSAGE('Hello %20s!', 'TEST');
NOTES:
Undocumented in 2012
Limited to 2044 characters
To escape the % sign, you need to double it.
If you are logging errors in extended events, calling FORMATMESSAGE comes up as a (harmless) error
take a look at xp_sprintf. example below.
DECLARE #ret_string varchar (255)
EXEC xp_sprintf #ret_string OUTPUT,
'INSERT INTO %s VALUES (%s, %s)', 'table1', '1', '2'
PRINT #ret_string
Result looks like this:
INSERT INTO table1 VALUES (1, 2)
Just found an issue with the max size (255 char limit) of the string with this so there is an alternative function you can use:
create function dbo.fnSprintf (#s varchar(MAX),
#params varchar(MAX), #separator char(1) = ',')
returns varchar(MAX)
as
begin
declare #p varchar(MAX)
declare #paramlen int
set #params = #params + #separator
set #paramlen = len(#params)
while not #params = ''
begin
set #p = left(#params+#separator, charindex(#separator, #params)-1)
set #s = STUFF(#s, charindex('%s', #s), 2, #p)
set #params = substring(#params, len(#p)+2, #paramlen)
end
return #s
end
To get the same result as above you call the function as follows:
print dbo.fnSprintf('INSERT INTO %s VALUES (%s, %s)', 'table1,1,2', default)
I have created a user defined function to mimic the string.format functionality.
You can use it.
stringformat-in-sql
UPDATE:
This version allows the user to change the delimitter.
-- DROP function will loose the security settings.
IF object_id('[dbo].[svfn_FormatString]') IS NOT NULL
DROP FUNCTION [dbo].[svfn_FormatString]
GO
CREATE FUNCTION [dbo].[svfn_FormatString]
(
#Format NVARCHAR(4000),
#Parameters NVARCHAR(4000),
#Delimiter CHAR(1) = ','
)
RETURNS NVARCHAR(MAX)
AS
BEGIN
/*
Name: [dbo].[svfn_FormatString]
Creation Date: 12/18/2020
Purpose: Returns the formatted string (Just like in C-Sharp)
Input Parameters: #Format = The string to be Formatted
#Parameters = The comma separated list of parameters
#Delimiter = The delimitter to be used in the formatting process
Format: #Format = N'Hi {0}, Welcome to our site {1}. Thank you {0}'
#Parameters = N'Karthik,google.com'
#Delimiter = ','
Examples:
SELECT dbo.svfn_FormatString(N'Hi {0}, Welcome to our site {1}. Thank you {0}', N'Karthik,google.com', default)
SELECT dbo.svfn_FormatString(N'Hi {0}, Welcome to our site {1}. Thank you {0}', N'Karthik;google.com', ';')
*/
DECLARE #Message NVARCHAR(400)
DECLARE #ParamTable TABLE ( Id INT IDENTITY(0,1), Paramter VARCHAR(1000))
SELECT #Message = #Format
;WITH CTE (StartPos, EndPos) AS
(
SELECT 1, CHARINDEX(#Delimiter, #Parameters)
UNION ALL
SELECT EndPos + (LEN(#Delimiter)), CHARINDEX(#Delimiter, #Parameters, EndPos + (LEN(#Delimiter)))
FROM CTE
WHERE EndPos > 0
)
INSERT INTO #ParamTable ( Paramter )
SELECT
[Id] = SUBSTRING(#Parameters, StartPos, CASE WHEN EndPos > 0 THEN EndPos - StartPos ELSE 4000 END )
FROM CTE
UPDATE #ParamTable
SET
#Message = REPLACE(#Message, '{'+ CONVERT(VARCHAR, Id) + '}', Paramter )
RETURN #Message
END
There is a way, but it has its limitations. You can use the FORMATMESSAGE() function. It allows you to format a string using formatting similar to the printf() function in C.
However, the biggest limitation is that it will only work with messages in the sys.messages table. Here's an article about it: microsoft_library_ms186788
It's kind of a shame there isn't an easier way to do this, because there are times when you want to format a string/varchar in the database. Hopefully you are only looking to format a string in a standard way and can use the sys.messages table.
Coincidentally, you could also use the RAISERROR() function with a very low severity, the documentation for raiseerror even mentions doing this, but the results are only printed. So you wouldn't be able to do anything with the resulting value (from what I understand).
Good luck!
Raw t-sql is limited to CHARINDEX(), PATINDEX(), REPLACE(), and SUBSTRING() for string manipulation. But with sql server 2005 and later you can set up user defined functions that run in .Net, which means setting up a string.format() UDF shouldn't be too tough.
I think there is small correction while calculating end position.
Here is correct function
**>>**IF OBJECT_ID( N'[dbo].[FormatString]', 'FN' ) IS NOT NULL
DROP FUNCTION [dbo].[FormatString]
GO
/***************************************************
Object Name : FormatString
Purpose : Returns the formatted string.
Original Author : Karthik D V http://stringformat-in-sql.blogspot.com/
Sample Call:
SELECT dbo.FormatString ( N'Format {0} {1} {2} {0}', N'1,2,3' )
*******************************************/
CREATE FUNCTION [dbo].[FormatString](
#Format NVARCHAR(4000) ,
#Parameters NVARCHAR(4000)
)
RETURNS NVARCHAR(4000)
AS
BEGIN
--DECLARE #Format NVARCHAR(4000), #Parameters NVARCHAR(4000) select #format='{0}{1}', #Parameters='hello,world'
DECLARE #Message NVARCHAR(400), #Delimiter CHAR(1)
DECLARE #ParamTable TABLE ( ID INT IDENTITY(0,1), Parameter VARCHAR(1000) )
Declare #startPos int, #endPos int
SELECT #Message = #Format, #Delimiter = ','**>>**
--handle first parameter
set #endPos=CHARINDEX(#Delimiter,#Parameters)
if (#endPos=0 and #Parameters is not null) --there is only one parameter
insert into #ParamTable (Parameter) values(#Parameters)
else begin
insert into #ParamTable (Parameter) select substring(#Parameters,0,#endPos)
end
while #endPos>0
Begin
--insert a row for each parameter in the
set #startPos = #endPos + LEN(#Delimiter)
set #endPos = CHARINDEX(#Delimiter,#Parameters, #startPos)
if (#endPos>0)
insert into #ParamTable (Parameter)
select substring(#Parameters,#startPos,#endPos - #startPos)
else
insert into #ParamTable (Parameter)
select substring(#Parameters,#startPos,4000)
End
UPDATE #ParamTable SET #Message =
REPLACE ( #Message, '{'+CONVERT(VARCHAR,ID) + '}', Parameter )
RETURN #Message
END
Go
grant execute,references on dbo.formatString to public
One more idea.
Although this is not a universal solution - it is simple and works, at least for me :)
For one placeholder {0}:
create function dbo.Format1
(
#String nvarchar(4000),
#Param0 sql_variant
)
returns nvarchar(4000)
as
begin
declare #Null nvarchar(4) = N'NULL';
return replace(#String, N'{0}', cast(isnull(#Param0, #Null) as nvarchar(4000)));
end
For two placeholders {0} and {1}:
create function dbo.Format2
(
#String nvarchar(4000),
#Param0 sql_variant,
#Param1 sql_variant
)
returns nvarchar(4000)
as
begin
declare #Null nvarchar(4) = N'NULL';
set #String = replace(#String, N'{0}', cast(isnull(#Param0, #Null) as nvarchar(4000)));
return replace(#String, N'{1}', cast(isnull(#Param1, #Null) as nvarchar(4000)));
end
For three placeholders {0}, {1} and {2}:
create function dbo.Format3
(
#String nvarchar(4000),
#Param0 sql_variant,
#Param1 sql_variant,
#Param2 sql_variant
)
returns nvarchar(4000)
as
begin
declare #Null nvarchar(4) = N'NULL';
set #String = replace(#String, N'{0}', cast(isnull(#Param0, #Null) as nvarchar(4000)));
set #String = replace(#String, N'{1}', cast(isnull(#Param1, #Null) as nvarchar(4000)));
return replace(#String, N'{2}', cast(isnull(#Param2, #Null) as nvarchar(4000)));
end
and so on...
Such an approach allows us to use these functions in SELECT statement and with parameters of nvarchar, number, bit and datetime datatypes.
For example:
declare #Param0 nvarchar(10) = N'IPSUM' ,
#Param1 int = 1234567 ,
#Param2 datetime2(0) = getdate();
select dbo.Format3(N'Lorem {0} dolor, {1} elit at {2}', #Param0, #Param1, #Param2);
Actually there is no built in function similar to string.Format function of .NET is available in SQL server.
There is a function FORMATMESSAGE() in SQL server but it mimics to printf() function of C not string.Format function of .NET.
SELECT FORMATMESSAGE('This is the %s and this is the %s.', 'first variable', 'second variable') AS Result
Here is my version. Can be extended to accommodate more number of parameters and can extend formatting based on type. Currently only date and datetime types are formatted.
Example:
select dbo.FormatString('some string %s some int %s date %s','"abcd"',100,cast(getdate() as date),DEFAULT,DEFAULT)
select dbo.FormatString('some string %s some int %s date time %s','"abcd"',100,getdate(),DEFAULT,DEFAULT)
Output:
some string "abcd" some int 100 date 29-Apr-2017
some string "abcd" some int 100 date time 29-Apr-2017 19:40
Functions:
create function dbo.FormatValue(#param sql_variant)
returns nvarchar(100)
begin
/*
Tejasvi Hegde, 29-April-2017
Can extend formatting here.
*/
declare #result nvarchar(100)
if (SQL_VARIANT_PROPERTY(#param,'BaseType') in ('date'))
begin
select #result = REPLACE(CONVERT(CHAR(11), #param, 106), ' ', '-')
end
else if (SQL_VARIANT_PROPERTY(#param,'BaseType') in ('datetime','datetime2'))
begin
select #result = REPLACE(CONVERT(CHAR(11), #param, 106), ' ', '-')+' '+CONVERT(VARCHAR(5),#param,108)
end
else
begin
select #result = cast(#param as nvarchar(100))
end
return #result
/*
BaseType:
bigint
binary
char
date
datetime
datetime2
datetimeoffset
decimal
float
int
money
nchar
numeric
nvarchar
real
smalldatetime
smallint
smallmoney
time
tinyint
uniqueidentifier
varbinary
varchar
*/
end;
create function dbo.FormatString(
#format nvarchar(4000)
,#param1 sql_variant = null
,#param2 sql_variant = null
,#param3 sql_variant = null
,#param4 sql_variant = null
,#param5 sql_variant = null
)
returns nvarchar(4000)
begin
/*
Tejasvi Hegde, 29-April-2017
select dbo.FormatString('some string value %s some int %s date %s','"abcd"',100,cast(getdate() as date),DEFAULT,DEFAULT)
select dbo.FormatString('some string value %s some int %s date time %s','"abcd"',100,getdate(),DEFAULT,DEFAULT)
*/
declare #result nvarchar(4000)
select #param1 = dbo.formatValue(#param1)
,#param2 = dbo.formatValue(#param2)
,#param3 = dbo.formatValue(#param3)
,#param4 = dbo.formatValue(#param4)
,#param5 = dbo.formatValue(#param5)
select #param2 = cast(#param2 as nvarchar)
EXEC xp_sprintf #result OUTPUT,#format , #param1, #param2, #param3, #param4, #param5
return #result
end;
here's what I found with my experiments using the built-in
FORMATMESSAGE() function
sp_addmessage #msgnum=50001,#severity=1,#msgText='Hello %s you are #%d',#replace='replace'
SELECT FORMATMESSAGE(50001, 'Table1', 5)
when you call up sp_addmessage, your message template gets stored into the system table master.dbo.sysmessages (verified on SQLServer 2000).
You must manage addition and removal of template strings from the table yourself, which is awkward if all you really want is output a quick message to the results screen.
The solution provided by Kathik DV, looks interesting but doesn't work with SQL Server 2000, so i altered it a bit, and this version should work with all versions of SQL Server:
IF OBJECT_ID( N'[dbo].[FormatString]', 'FN' ) IS NOT NULL
DROP FUNCTION [dbo].[FormatString]
GO
/***************************************************
Object Name : FormatString
Purpose : Returns the formatted string.
Original Author : Karthik D V http://stringformat-in-sql.blogspot.com/
Sample Call:
SELECT dbo.FormatString ( N'Format {0} {1} {2} {0}', N'1,2,3' )
*******************************************/
CREATE FUNCTION [dbo].[FormatString](
#Format NVARCHAR(4000) ,
#Parameters NVARCHAR(4000)
)
RETURNS NVARCHAR(4000)
AS
BEGIN
--DECLARE #Format NVARCHAR(4000), #Parameters NVARCHAR(4000) select #format='{0}{1}', #Parameters='hello,world'
DECLARE #Message NVARCHAR(400), #Delimiter CHAR(1)
DECLARE #ParamTable TABLE ( ID INT IDENTITY(0,1), Parameter VARCHAR(1000) )
Declare #startPos int, #endPos int
SELECT #Message = #Format, #Delimiter = ','
--handle first parameter
set #endPos=CHARINDEX(#Delimiter,#Parameters)
if (#endPos=0 and #Parameters is not null) --there is only one parameter
insert into #ParamTable (Parameter) values(#Parameters)
else begin
insert into #ParamTable (Parameter) select substring(#Parameters,0,#endPos)
end
while #endPos>0
Begin
--insert a row for each parameter in the
set #startPos = #endPos + LEN(#Delimiter)
set #endPos = CHARINDEX(#Delimiter,#Parameters, #startPos)
if (#endPos>0)
insert into #ParamTable (Parameter) select substring(#Parameters,#startPos,#endPos)
else
insert into #ParamTable (Parameter) select substring(#Parameters,#startPos,4000)
End
UPDATE #ParamTable SET #Message = REPLACE ( #Message, '{'+CONVERT(VARCHAR,ID) + '}', Parameter )
RETURN #Message
END
Go
grant execute,references on dbo.formatString to public
Usage:
print dbo.formatString('hello {0}... you are {1}','world,good')
--result: hello world... you are good
At the moment this doesn't really exist (although you can of course write your own). There is an open connect bug for it: https://connect.microsoft.com/SQLServer/Feedback/Details/3130221, which as of this writing has just 1 vote.
Not exactly, but I would check out some of the articles on string handling (amongst other things) by "Phil Factor" (geddit?) on Simple Talk.
this is bad approach. you should work with assembly dll's, in which will do the same for you with better performance.