The docs here say that table location has a field called featureTypes. But in the most recent version of the dumps, the column seems to be named featureType. Probably the docs should updated accordingly? Or should the csv files should be changed, since it is comma separated and Places does have a column called featureTypes.

Comma separated list like

dare:importance=2,dare:symbol=settlement,...

Since dumps come from the catalog, this will require cataloging of the precision value. Raises the question of whether positional accuracy precision can be merged with our other sense of precise vs rough.

Some records in the dump have new lines inside fields. This is ok when you load the whole CSV as a spreadsheet in order to work on it, but it means that lots of command-line tools will have trouble with the file. For example, I was using grep to select entries that had a featureType of temple and the new lines meant that those entries got split mid-field.

For example, pleiades place 786059 has a new line in the description field, which shows up as \r\n in the json.

Any way to represent the new line in a different way inside fields in the dump to allow use of such tools? (It's not clear to me that the new lines are needed/desirable anyway.)

Hi

Can you please indicate how each of the 3 data sets are to be joined with each other - the column names - and whether it is normal to expect duplicate records if the data sets are merged?

I was merging the location and names data sets to places data set using pid = path, but I end up getting duplicate rows.

Thanks a lot

Could you add the references to the .csv dump?