Skip to content

Instantly share code, notes, and snippets.

@simonmichael
Last active December 2, 2021 20:53
Show Gist options
  • Star 0 You must be signed in to star a gist
  • Fork 0 You must be signed in to fork a gist
  • Save simonmichael/feee7fc1567c5f10355f952b87bd3090 to your computer and use it in GitHub Desktop.
Save simonmichael/feee7fc1567c5f10355f952b87bd3090 to your computer and use it in GitHub Desktop.
hledger-number - extract one number from hledger
#!/usr/bin/env bash
# hledger-number - try to extract just one single machine-readable number from hledger:
# the grand total of `hledger balance` run with any provided arguments.
# Requires hledger 1.24 or newer (the December 2021 release, see https://hledger.org/download.html )
# https://gist.github.com/simonmichael/feee7fc1567c5f10355f952b87bd3090
VALUATION_COMMODITY="$"
hledger bal -0 -N -X "$VALUATION_COMMODITY" --infer-market-prices -c "$VALUATION_COMMODITY 1000" --layout=bare "$@" | awk '{print $1}'
# Tired of complex financial reports ? This is a silly but fun and
# occasionally useful script showing how to get "one number" semi-robustly
# without writing a haskell script. With this installed in PATH, then eg:
#
# $ hledger number lodging
# 200
#
# Explanation:
# -0 (--depth 0) hides all but top-level accounts
# -N (--no-total) hides the totals line
# -X COMM (--value=end,COMM) converts to a single commodity if possible (needs at least one suitable P market price declaration)
# --infer-market-prices guesses P price from conversion transactions if necessary
# -c (--commodity-style) sets a predictable number format free of thousands separators
# --layout=bare moves the commodity symbol away from the number
# awk discards all but the number
#
# A note: If you're new to hledger/PTA and thinking it shouldn't be this hard..
# well of course normally when a hledger user wants to see one number, they run
# a much simpler command like `hledger bal lodging` or `hledger reg lodging`.
# This script is exploring how to anticipate and neutralise all variations
# caused by account structure, number formatting, mulltiple commodities, etc.
# so as to *reliably* produce *one machine readable number* from highly diverse
# data. As we figure out good approaches we'll build them in so that eg
# producing data for charting is as easy as it can be, while still being
# general.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment