Disjoint date periods in a single report are not well supported by most accounting software, see recent hledger issue. Nice solution!

Update: Here's the "Why not Lemmy" bit from my post above:

Lemmy is like Reddit but decentralised and run on a non-profit basis by many collaborators. This is great, and a much better fit, but it has problems of its own: it doesn’t work with my main browser Safari, the RSS feed is difficult to announce in chat, it is technically complex and immature, and it is not sufficiently different from Reddit to motivate reddit users to move there. I’m grateful it exists but at present I find it underwhelming for our needs.

Because of this, and positive early response to the experiment (including from the Reddit and Lemmy mods), and my personal readiness to try a more featureful setup, I'll keep going with this. I invite you all to visit, join, and/or cross-post there to get more eyeballs and answers:

https://forum.plaintextaccounting.org

Yes, no real benefit that I see, especially since you can generate them on the fly with --infer-equity when you want to see a perfectly balanced bse report. Though for some complex transactions inferring won't work and you'd have to manually add the equity postings to those.

PS one big win for me has been switching to just for managing my hledger-related scripts efficiently. Now it's cheaper for me to create new custom reports and to remember and reuse them.

I agree entirely. This is a weak area for myself and probably many PTA-ers... after learning the tools, managing your files, setting up a chart of accounts, tending your CSV rules, entering/downloading/converting/importing transactions, reconciling... there is likely not a lot of time or mental bandwidth left for higher-level analysis !

I feel making every step easier along the way helps to increase the chance of getting there. Another approach is to bypass some of the earlier steps; the only way that comes to mind is to periodically record balances (ie, large rollup "transactions") rather than individual transactions. I haven't really tried this myself nor seen it much written about, but that's the approach taken by the recent gainstrack app. It also suggests using a higher-level DSL for recording transactions, when you do record them.

I believe what's needed is

1. more discussion and sharing on this topic, like yours
2. some solid FAQ answers & how-tos on hledger.org and/or plaintextaccounting.org, describing how to achieve such reports with current tools
3. leading to some easier built-in reports & docs.

I don't really know what reports would be useful to me beyond the basics. It sounds like you have identified two good candidates. https://hledger.org/scripts.html#hledger-combine-balances is a script for comparing two periods. I also always want an easy way to make simple charts, and https://hledger.org/scripts.html#hledger-bar is the way I currently do that. There might be other scripts of interest on this page also (hledger-plot, hledger-lots..).

hledger is --alias expenses:job:tax=income:job:main should do it.

You can use an if rule with a regular expression pattern to match on positive/negative. Eg

if %theamountfield /^-/

But that doesn't sound helpful here since as you say they are both debits.

The payments from the bank should have fairly consistent descriptions, can't you filter on that ?