Hello James, >> I think the usual practice is to use BLOB types to store data that is not a character stream. So you case is probably pretty rare <<
admittedly - if the fields had been left as clob fields, then all would have been well. the change to informix Text blobs was driven by the need to use the informix dbload utility, to push data in to the target table before using the DIH to pull data from the target table in to the core. >> If casting solves the issue, then why not? << ok - i will concede this point - but i am interested in "why" ClobTransformer _needs_ the cast to work in the first place. >> Then again if CLOBTransformer was changed to handle BLOBs also, I do not see the harm << if possible - i would like to understand more about ClobTransformer and what would be needed to make that change. >> But I would think it would be a much more common case that users would be putting binary-format documents in BLOBs then feeding them to tika or something to extract the text. << i am not sure - maybe. at SHC (Sears) the data being stored in these two columns is a large JSON blob. when a query is performed, the JSON blob is parsed and used as needed. thanks again for the discussion and education. mark -- View this message in context: http://lucene.472066.n3.nabble.com/have-developer-question-about-ClobTransformer-and-DIH-tp4064256p4064289.html Sent from the Lucene - Java Developer mailing list archive at Nabble.com. --------------------------------------------------------------------- To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
