Release announcements ?

User f67d4188b6

17-12-2012 15:35:29

It would be save a lot of time it there would be release announcements of the Infocom Knime nodes.


At this moment Knime 2.7 has been released, but we can not upgrade until the infocom nodes for Knime 2.7 have been released as well.

User 5458277630

18-12-2012 02:32:47

I'm sorry to have kept you waiting.
We will release new version at mid-Jan, 2013.
However, the latest version (2.6.3.v0135) can work in KNIME 2.7.

Best,
Taka

User f67d4188b6

19-12-2012 09:17:27

Unfortunately upgrading to knime 2.7 while using the 2.6 nodes breaks some functionality.


 


When
reading an instantjchem table using the normal database reader node it now
puts the cd_structure column in a blob column, for which knime 2.7 now
has native support. But the Advanced Molconverter does not know about
this new datatype yet, and thus fails to see it entirely.


Converting
the column to a String column (using the Column Rename node) clips off
the information of moderately large molecules, thus this is not an
option either.


Only converting
the column to a text column using a SQL Convert ( SELECT CONVERT(cd_structure USING utf8) FROM MOLECULES ) does the trick, but as this needs to be changed in all workflows that use the database reader, it is not a workaround that we will use just yet.


 


OpenBabel seems to have the same issue with blobs by the way.

User 5458277630

19-12-2012 12:01:56

Hi,

I will enhance the MolConverter node to support BLOB type.

Meanwhile, would it be possible that you use the "Binary objects to strings" node instead of the "Column Rename" node?

I'm sorry for the inconvenience I've caused you.

Best,
Taka

User f67d4188b6

19-12-2012 12:12:31

This seems to work very well, thank you.

User f67d4188b6

19-03-2013 16:34:46

The recent update to Instant JChem version 5.12 again breaks the JChem Search and other nodes that directly access the database.


Ofcourse this can be circumvented using the database reader and converting the structures (first from blob to string, then from string to marvin structure), but this again points out the problem of the nodes not being in-sync with the chemaxon releases.


 


It would be very helpfull if this organizational problem could be adressed.