Log file Viewer
#mongodb logs for Wednesday the 8th of May, 2019
[12:59:53] <syntaxfree> eh. I'm inheriting this <frolicked up> project from a technical partner who had a mental breakdown and left slamming the door.
[13:01:42] <syntaxfree> I don t expect you to know (or not to know either) most of the names in this since it's python not mongo per se, but it uses geopandas, a dataframe (table)-like structure that has a column with geometry objects (multi/polygons, etc.). which are then "pickled" to disk.
[13:02:39] <syntaxfree> as a result of his frolicked up copy-paste coding style, there are multiple copies of these files because he addresses them by relative paths and has different modules doing different things. it's a mess. hundreds of shapefiles and no metadata.
[13:04:10] <syntaxfree> I'm considering starting a mongodb store to hold the raw geometry assets in this thing I just found on google called geojson. Apparently a standard serialization format for geometric objects that looks a lot like shapely objects in first place.
[13:06:27] <syntaxfree> does this make sense at all? seen from the top of a building, is the cost of calling a mongodb database a hundred times in half a second comparable or cheaper than opening files from disk (which maybe cacheable on memory in some way that currently escapes me, like old style ramdisks)?
[13:07:46] <syntaxfree> it helps that we have someone who's already deploying something like a crm database in mongo. But I don't know mongo per se, I barely know sql. I was hoping I would have articulated a clear question by now but didn't.
[13:13:24] <syntaxfree> aaand some additional google has made this rant irrelevant. I wish I could delete messages on irc like you can on slack and telegram.
[18:31:02] <Banana51> I will not be inserting new data frequently, I just need to optimize for very fast reads