bent.s.lund@gmail.com: It currently only casts empty strings to when they are used as Node IDs, which is arguably the most important case.

We're working separately on casting all strings to nulls to reduce overall property store size from a load created by data importer. In hindsight I probably shouldn't have merged the two posts, your request is equally valid.

Gregory King: Hi, 0.1.2 Beta seems to still create properties for empty columns in a .tsv-file. If I understand correctly, empty string should not result in properties being created. I can provide test-data if you want to investigate this

Data Importer has now been released with support for "Node ID exclude lists". If you need to treat more values than the empty string as null, you can find the setting in graph model settings.

Hi, this is being worked on at the moment and will be available in the next release of Data Importer. As well as treating empty strings for IDs as null by default, we're looking to add support for users to define specific strings to be treated as null, eg "undefined", "unknown" in settings.