Bulk load of address data fails

I am trying to load the region/district/ward/post code/village name/lat and long for about 15000 villages in Tanzania (by the way I am happy to share the data with anyone that needs it). However the load fails after about 3000 records. I have broken the data into 15 sets of about 1000 records which appear to load OK.

This seems consistent with some sort of data volume/time out restriction. Can anyone advise how to push up the boundaries?

Looks like a related discussion: Trying to build a comprehensive address hierarchy for India

I am not sure on how to push the performance boundaries for this. You could try out the scripts suggested in the above thread.

Hi @mafrica,

You could use the upcoming address configuration loading feature of Address Hierarchy. It loads the address template, the address levels and the address hierarchy (the CSV) upon starting the module. This will be released as part of 2.11. For now you would have to use our fork and clone & build the ADDR-109 branch. The PR is still being reviewed but hopefully this should be merged in soon. Would be great to get your feedback on this.

It’s pretty straightforward, the address configuration must be copied into configuration/addresshierarchy … and that’s it, just let the module start. Here is a configuration example, you can ignore the .properties files for your sake. This example loads in 16,000 entries with no problem.

Segmentation seems to have got me by for now. My main interest was to alert others to the problem which seems to have been done. Happy to try when 2.11 is released.