|
Hi All,
May be wrong place to say but we want to create functions on FIREBIRD and want to use in SELECT statements. Like this, CREATE MY_FUNC_CALC_AGE(prmBirthDay TIMESTAMP) RETURNS RET_AGE INTEGER BEGIN RET_AGE = DATEDIFF(YEAR FROM prmBirthDay TO CURRENT_DATE) ; SUSPEND ; END SELECT NAME, SIRNAME, MY_FUNC_CALC_AGE( BIRTHDAY) FROM MY_TABLE When this issue will planned to solve ? Thank you and best regards Adnan It's implemented in FB v3.0.
Is there any documentation available on this, as to how one would go about creating an external aggregate function in FB 3.0?
It's simple PSQL functions which is implemented. Aggregate function is not nor will be in FB 3.0.
Sadly. is there any plan to support it on a later version? It's known which version?
What's the rationale for this? Most other serious RDBMSes have a way to define external aggregates. It's something that's very noticeably lacking from Firebird.
Probably in the next major version.
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||
 |
|
New Feature
Open
Minor
Aggregate UDF
mysql support this and maybe other databases.
http://books.google.com/books?id=kVB1wiF87ooC&pg=PA265&lpg=PA265&dq=mysql+udf+aggregate+example&source=bl&ots=HOlP34LbdM&sig=Fcj_7w5NycrtKhxpUbAREpMQ5K0&hl=en&ei=saTdSub8OIGUjAe838xt&sa=X&oi=book_result&ct=result&resnum=5&ved=0CBkQ6AEwBA#v=onepage&q=mysql%20udf%20aggregate%20example&f=false
why is this important?
if I need functions which contains few aggregate function then the best way for this problem is aggregate UDF.
firebird c-api does not support this and I don't know why?
many special functions which people need can't be write in sql query.
can be but with two or more stored procedures.
for example consider standard deviation
http://en.wikipedia.org/wiki/Standard_deviation
first query is avg(x) on stored procedures.
input parameter is one column-> x.
then must using cursor and for each x calculate
(x-avg(x))^2 as y
then
must run query
sqrt(sum(y)/count(x))
I know that somebody will say that database is not for processing data.
cursor is very slow, ETL (SISS) is fast but for processing data is slow.
ETL(SISS) is only for transform and migrate data.
application for statistics is also not very good.
what users want is only enable aggregate UDF in c-api and nothing else.