I am currently working on preparing data for plotting a Linemap. Natively we have the data in the following format :
id | lat_start | lon_start | lat_end | lon_end
Ideally, I would like to solve this within the OmniSciDB. I saw in this thread and in the documentation that it is possible to convert lon and lat to a point with the ST_Point function. However, there does not seem to be a similar function for loading in
LINESTRINGS. Is this correct?
A second approach I thought of was to concat values together to hack myself towards a WKT representation could then maybe be used in the ST_GeomFromText function. However, this was not possible as my longitude and latitude data is stored numerically and OmniSciDB does not seem to like being casted to TEXTs (as also established in this thread).
I then tried to use IBIS but realized that even though the interface is python the background code is just being compiled by the DB and therefore the casting limitations are the same or?
Lastly, as the table is not especially large I could theoretically just download it, do the transformation in python and then upload it again. However, it seems that loading pandas dataframes through ibis is not yet supported in the create_table() function or?
I would very much appreciate any suggestions on how to approach or circuvemt these problems.
Please sign in to leave a comment.