These scripts are pretty complicated. This isn't ideal, but it's
necessary: InfluxDB 1 doesn't have a very good query language, so
hledger-to-influxdb needs to do a lot of work to mangle the data
into a format I can do what I want with.
But every time I want to do something new with the data, I need to go change the script. And occasionally the script fails because it times out uploading all the data to InfluxDB.
I realised I could do much less pre-processing, dramatically simplifying the script and making it more reliable, if I migrated to a timeseries database with a more powerful query language. The obvious choices were InfluxDB 2 and Prometheus.
InfluxDB 2 is very different, and I don't use it at work or in any other projects, I would be learning it just to create one dashboard. That wasn't very appealing.
Prometheus is something I already use for monitoring my personal things, and I use it at work too. But I couldn't figure out how to get my historic data into it.
Now I've solved that problem. I found promscale, a Prometheus-compatible timeseries database which supports bulk importing old data.
So these scripts are no longer used. Go see [my new, much
simpler, script][] which is written in Python and uses hledger print -O csv rather than interfacing with the hledger library.