Update the dump

[SimonVilleneuvewp]

Hi,

First, thank you so much for this tool. We have installed it and it work perfectly ! I'm really happy to have access to my own Wikidata query service.

I want to know if we must do all the work again with each new dump or if we can only do a kind of "update" ?

[joka921]

Hi,
Thanks for your feedback and for considering QLever.
Currently QLever does not support updating of the Index, but you have to fully rebuild it every time.
We have already discussed the general approach on how this could be implemented in QLever (see here for a technical discussion, but this is currently not a high priority for us (we currently focus on missing features and reduction of RAM and harddisk usage).

Note that this is in practice currently not a big limitation especially for Wikidata, since they only release weekly full dumps anyway and it then takes about a day to rebuild the QLever index (we have an automated pipeline that in fact builds a new Wikidata index each week to keep our demo instance up to date).

Most other large knowledge graphs we are aware of are similarly "lazy", meaning that the update frequency is much slower than the time needed for a full QLever index build.

Please let us know about any other questions and findings you encounter when using QLever for Wikidata.