When generating the schema generation script, the terminate statement ";" is generated correctly when using a h2 database, but no terminate statement is generated for oracle. (eclipselink.ddlgen-terminate-statements is set to true)
Is there any way to make eclipselink generate ";"? Really want to avoid going through the script afterward and add ";" manually...
I found a workaround here shortly after having posted the question.... Basically,
Create a subclass of OraclePlatform
set the protected field storedProcedureTerminationToken to ;
Use its qualified name for the property eclipselink.target-database
Related
I have to remove the accents from the person's name, but I cannot apply the function in Talend while it works in SQL oracle.
this query works in my tDBInput component :
"SELECT '"+((String)globalMap.get("copyOfSORTIE.NOM"))+"' as nom_nom_compl,
'"+((String)globalMap.get("copyOfSORTIE.ENTETE"))+"' entete
FROM DUAL"
However, when I want to add the convert function, it doesn't work
this query does not work :
"SELECT '"+((String)globalMap.get(CONVERT("copyOfSORTIE.NOM",'US7ASCII')))+"' as nom_nom_compl,
'"+((String)globalMap.get("copyOfSORTIE.ENTETE"))+"' entete
FROM DUAL"
In my talend :
I am getting this error
What is the syntax for it to work?
Thank you!
Two things there :
I don't know the CONVERT method, but I can see that you are applying it to the key of your globalMap variable , and not the value (as if you wanted to convert "myKey" and not "myValue" which is attached to the key). Are you sure this is what you want to achieve ? if not, the syntax should be something similar to "SELECT CONVERT('"+((String)globalMap.get("copyOfSORTIE.NOM"))+"','US7ASCII') "
A useful java method implemented in talend is TalendString.removeAccents("") that you can apply directly on your talend variable, thus not using a SQL method.
i need to do a SQL simulator on Prolog. i need to simulate some functions like create_table, insert, delete, update_table, select_table, drop_table, show_table, etc. I am learning how to do asserts and retracts but im getting some errors in my first function create_table(N,A) where N is the name of the table and A a list with the atributtes
An example is create_table("student",["name","lastname","age"]). This will create my table named "student" with the atributes ["name","lastname","age"].
I was searching how to work with assert and i see i need to do dynamic before making assert, then my code is.
len([],0).
len([_|T],N) :- len(T,X), N is X+1.
create_table(_, []):-!.
create_table(N, Atributos):- len(Atributos, L), dynamic N/L, assert(N(Atributos)).
But i get this error :7: Syntax error: Operator expected on this line
create_table(N, Atributos):- len(Atributos, L), dynamic N/L, assert(N(Atributos)).
What im doing wrong? excuse me, i dont speak good english.
From the error message, seems you're using SWI-Prolog....
Your code could be (note that len/2 can be replaced by the builtin length/2)
create_table(N, Atributos):-
length(Atributos, L),
dynamic(N/L),
T =.. [N|Atributos], % this missing 'constructor' was the cause of the error message
assert(T). % this is problematic
There is an important 'design' problem: the CREATE TABLE SQL statement works at metadata level.
If you do, for instance,
?- assertz(student('Juan', 'Figueira', 20)).
pretending that the predicate student/3 holds table data, you're overlapping data and metadata
using dynamic/1 and assert is a tricky non-logical aspect of Prolog, and dynamically creating dynamic predicates is really unusual. Fundamentally you cannot have a Prolog query with the predicate name as a variable e.g.
query(N,X) :- N=student, N(X).
My suggestion is you remove a layer of complexity and have one dynamic predicate 'table', and assert your SQL tables as new 'table' clauses i.e.
:- dynamic table/2.
:- assertz(table(student,['Joe','Young',18])).
query(N,X) :- table(N,X).
:- query(student,[Name,Lastname,Age]).
I've a problem with escaping colons and dashes in SQL queries when I use dbFit with Fitnesse.
Such statement doesn't work:
!|Query|select to_char(my_birthday,'YYYY-MM-DD HH24:MI:SI') from family|
I need to replace colons and dashes with some other acceptable characters, ex.
!|Query|select to_char(my_birthday,'YYYY_MM_DD HH24_MI_SI') from family|
Do you know how to solve it properly without using the 2nd approach ?
Cheers,
foxrafi
I think this is what you need. From http://dbfit.github.io/dbfit/docs/reference.html
Avoiding parameter mapping
If you want to prevent DbFit from mapping parameters to bind variables
(eg to execute a stored procedure definition that contains the #
symbol in Sql Server), disable bind symbols option before running the
query.
|set option|bind symbols|false|
|execute| insert into users (name, username) values ('#hey','uuu')|
|query|select * from users| |name|username| |#hey|uuu|
Remember to re-enable the option after the query is executed. You can
use the same trick with the Execute command.
In addition to Mike's answer, you can also solve this by using bind parameters. This is useful if you have to use bind parameters at other places in the same query.
!|DatabaseEnvironment|ORACLE|
|Connect|${HOSTNAME}|${USERNAME}|${PASSWORD}|
!|Query|!-select current_timestamp ts from dual-!|
|ts?|
|>>my_birthday_ts|
|set parameter|my_birthday_ts| <<my_birthday_ts|
#Set bind parameter :MI to string ':MI'
|set parameter|MI|:MI|
# and do it in the same way with :SS.
|set parameter|SS|:SS|
!|Query|!-select to_char(:my_birthday_ts, 'YYYY-MM-DD HH24'|| :MI || :SS) bds from dual-!|
|bds?|
|>>birthday_string|
Note that you have to use !- -! around your query, otherwise FitNesse will expand the concatenation operator to table cells. The main drawback of this manner is that you cannot use ordinary FitNesse variables (${varname}) in the query.
!|DatabaseEnvironment|ORACLE|
|Connect|${HOSTNAME}|${USERNAME}|${PASSWORD}|
!|Query|!-select current_timestamp ts from dual-!|
|ts?|
|>>my_birthday_ts|
!|Query|!-select to_char(:my_birthday_ts, 'YYYY-MM-DD HH24:'||'MI:'||'SS) bds from dual-!|
|bds?|
|>>birthday_string|
If I were retrieving the data I wanted from a plain sql query, the following would suffice:
select * from stvterm where stvterm_code > TT_STUDENT.STU_GENERAL.F_Get_Current_term()
I have a grails domain set up correctly for this table, and I can run the following code successfully:
def a = SaturnStvterm.findAll("from SaturnStvterm as s where id > 201797") as JSON
a.render(response)
return false
In other words, I can hardcode in the results from the Oracle function and have the HQL run correctly, but it chokes any way that I can figure to try it with the function. I have read through some of the documentation on Hibernate about using procs and functions, but I'm having trouble making much sense of it. Can anyone give me a hint as to the proper way to handle this?
Also, since I think it is probably relevant, there aren't any synonyms in place that would allow the function to be called without qualifying it as schema.package.function(). I'm sure that'll make things more difficult. This is all for Grails 1.3.7, though I could use a later version if needed.
To call a function in HQL, the SQL dialect must be aware of it. You can add your function at runtime in BootStrap.groovy like this:
import org.hibernate.dialect.function.SQLFunctionTemplate
import org.hibernate.Hibernate
def dialect = applicationContext.sessionFactory.dialect
def getCurrentTerm = new SQLFunctionTemplate(Hibernate.INTEGER, "TT_STUDENT.STU_GENERAL.F_Get_Current_term()")
dialect.registerFunction('F_Get_Current_term', getCurrentTerm)
Once registered, you should be able to call the function in your queries:
def a = SaturnStvterm.findAll("from SaturnStvterm as s where id > TT_STUDENT.STU_GENERAL.F_Get_Current_term()")
I'm converting some RDO code to ODBC Provider code in .NET.
The problem is parameter names were not specified in the orignal code, but param values were retrieved by parameter name after the command was executed.
Is there anyway to have parameter names populated by the provider once the command is executed so calling code can access params by name.
Let me show you an example of the declaration of param and accessing of it.
With rdqryClntBasic
.Parameters.Add(.CreateParameter) : .Parameters(0).Direction = ParameterDirection.Input
.Parameters(0).DbType = DbType.String
.Parameters(0).Value = sClntProdCd
End With
.EffectiveDate = ToDate(rdqryClntBasic.Parameters("dtEffDt").Value)
You can now see how this "used to work in RDO/VB". For some reason it would accept this and know what the param names were after execution. I imagine it had to do another round trip to the db to get this info.
Is there anyway to mimic this behaviour in .NET for ODBC Provider (using Oracle)? Or am I stuck manually specifying the param names in the code (I understand this is the better option, but wondering what the alternative is to match the original code as closely as possible).
No, parameters in ODBC are positional not by name.