Configuring Documentum file extractor meta-data

@lhcy at the moment there is no way to filter/select metadata fields. If you would like to have this as a feature, I can convert this question into a product idea.

Hi @Mithila Jayalath , that would be appreciated! Flexbility to remove trivial meta-data will make Cognite Search easier to use. 

Hi @lhcy,

Am I right in describing the problem you’re seeing is the existence of too many metadata keys (fields) in the data from your Documentum source making it difficult to search for documentum data in CDF?

Should the requested capability exist at some point, are there any concerns of perhaps being too aggressive when selecting the metadata fields to exclude and what that may mean for the user experience? (I do realize that may be remedied by re-ingesting with a less stringent exclusion list/filter). 

Is the problem solely that the “full list of metadata” from the source system represents a confusing amount of data, or is there specific data you think it makes sense to exclude? Maybe by default?

Hi @Thomas Sjølshagen, the ideal feature is giving us ‘flexibility’ to select metadata during ingestion process - so that we can add / reduce metadata based on user feedback (this could be an iterative process initially). 

Another issue we face with Documentum metadata (other than having too many making it hard to Search), is Documentum Extractor runs failing due to certain files exceeding allowable metadata limit, e.g. for us culprits are the <tag_num> and <links> metadata. 

Being able to select only relevant metadata might solve, or at least improve, some of our memory issues. When retrieving a large amount of assets, where we only need 2 metadata fields but have to retrieve all 30+, this adds up to quite a bit when we are talking about 500k+ assets (and no, we can’t filter to limit the number, we need the full context).

Basically, sending in the request with a list of keys specifying the relevant metadata fields.