Hi,
I'm trying to define an aggregate function in MonetDB to concatenate
rows in select group by (GROUP_CONCAT in MySQL). Postgres does not have
this but there is an example to define this function.
I'm starting to use monetdb from mysql thanks to MonetDB performance
for large data sets. For some small tables, we have the feature to
allow users to store small amount of custom data in rows and have a web
service to convert rows into columns. We have this feature because we
can do it in MySQL and Postgres using GROUP_CONCAT. I'm working on an
open source breeding system which stores breeding data in MySQL,
environment data in Postgres and now marker data in MonetDB in
Australia. The system is called KDDart.
Going back to my problem is that there is the same question with
GROUP_CONCAT about 3 years ago. I have been googling and trying based
on the information available for 2 days now. I have learnt how to
define a MAL function and make it available in SQL mode by saving it in
the autoload. However, when it comes to define and write a MAL
aggregate function. I'm unable to find any step-by-step example. I
could not start successfully with a simple aggregate solution.
I google further that we a documentation on monetdb website about
defining an aggregate function in R. I installed R and recompile my
MonetDB source (MonetDB-11.19.9) on my openSuSe 12.2 with a
pre-compiled R (R-base-2.15.1-1.1.4.x86_64). I followed the instruction
to enable rintegration and compile successfully. R integration is
loaded with mserver5 instruction. But when using monetdbd it says 'set:
no such property: embedr'. I did bit of C debugging but I only got
Segmentation fault.
I don't know what you could do to help me out. I wish I could write the
GROUP_CONCAT MAL function and being able to define it permanently in
.mal file and being able to link it to the SQL.
I hope you can help me out.
Thanks in advance.
Puthick