search

bjornd / jvectormap

jvectormap.com
GNU Affero General Public License v3.0
1.38k stars 431 forks source link

Entries from shape file missing after converting #449

Open jpcwilliams opened 5 years ago

jpcwilliams commented 5 years ago

Hi, We are trying to convert shape files provided by Landgate and have found that a large number of entries are missing from the converted file. Can't seem to find anything common between the missing entries.

The input_file.json is [{"name": "read_data", "file_name": "/tmp/tmp-29310yOFosO3W0Pvp.1/Cadastre .shp"}, { "name": "write_data", "format": "jvectormap", "file_name":"/tmp/tmp-29310MFms0TQWd48e.tmp.1", "params": {"code_field": "polygon_nu", "name_field": "polygon_nu", "name": "au-wa-xxx"}}]

In the dbf file polygon_nu in unique but when searching for the numbers from that file in the output file we can find only about half of them. Missing entries and included entries seem to be almost exactly the same. Missing entries are at random throughout the dbf file so it's not reading part way then stopping. No errors reported by the converter. Would appreciate any ideas on how to track down the issues

jpcwilliams commented 5 years ago

Found the issue. buffer_distance was causing a lot of entries to be missed. Changing this value to 0 and everything is now included but end file size is huge.