in our Oracle DB we have a column of type
varray (3800) of BINARY_FLOAT
which I would like to map to a KNIME collection type.
In Knime analytics platform 4.1.3 (also tried in the nightly 2.7.2020) I’m using the (non-legacy) DB Connector with a JDBC8-19 driver. In the input type mapping this column is shown as “array” and mapped to “string->string” which shoes me a hash, but not the data (which would be a float array).
I cannot find a mapping to a collection.
Is there a way to map this float array to a KNIME collection type? Or any other recommendation how to handle it?
These float array values represent a kind of compound descriptor which we would like to use to predict compound properties. So for each compound we would like to use this vector as input for a machine learning algorithm to train a model and do predicions.
@Jeany I guess this might be a common scenario how KNIME is used in Pharma and might help others as well, so any help here is highly appreciated!