I did try with din namespace and that didn't seem to make any difference. Since the PK is a composite in my case, just specifying the bib_id was throwing an exception stating "could not find the matching pk column" or something to that effect. Although I realize the use cases for using one or the other strategy, since in my case I am only updating one row at a time or perhaps 1K at most, I should be able to use the "all in one query" strategy for the imports rather than having three separate ones.
Thanks On Jun 2, 2013, at 1:25 PM, Shawn Heisey <s...@elyograg.org> wrote: > On 6/2/2013 10:11 AM, PeriS wrote: >> I found using the strategy mentioned at >> http://wiki.apache.org/solr/DataImportHandlerDeltaQueryViaFullImport, it >> works for me. Not sure what the difference is between this one and writing >> individual queries for fetching the IDs first and then getting the data; I >> mean I know the difference, but not sure why the previous way doesn't work. > > Glad you found a solution. As for the initial problem: > > I am guessing that the pk value in the config file should be just > "bib_id". That value is only used within dataimporter, it has no > bearing on the Solr index. I think it also has nothing to do with the > interaction with the database. As far as I know, it only gets used to > validate SQL input data to make sure it is sane, especially with delta > imports. > > Also, there might be a problem with the dataimporter namespace for > subsitution variables, that has migrated to dih in newer versions, so > you need things like this: > > ${dih.delta.bib_id} > ${dih.last_index_time} > > I have seen problems where the dataimporter namespace hasn't worked, dih > is much safer. > > Thanks, > Shawn > > > > >