I discovered John Wiegley's Ledger in 2006, and was very happy to find this efficient command-line reporting tool with a transparent data format.
Initially, I used it to generate time reports for my job. Before long I wanted that to work differently - splitting sessions at day boundaries, reporting in hours, etc. John had got busy elsewhere and the Ledger project now stalled, with unfixed bugs, wrong documentation and a confusing release situation persisting for a long time. I did what I could to help build momentum, reporting bugs, supporting newcomers, and contributing a new domain and website. But, I didn't want to spend time learning C++.
I was learning Haskell, which I did want to spend time in. I felt Ledger could be implemented well and, in the long run, more efficiently in that language, which has some compelling advantages such as lower maintenance costs. I urgently needed a reliable accounting tool that I enjoyed using. I also wanted to see what I could do to reduce roadbumps and confusion for newcomers.
I couldn't expect John to start over - at that time he was not the Haskell fan he is now! So in 2007 I began experimenting. I built a toy parser in a few different languages, and it was easiest in Haskell. I kept tinkering. Goals included:
Before too long I had a tool that was useful to me. With Ledger still installed, and by maintaining high compatibility, I now had two tools with different strengths, each providing a comparison for the other in case of confusion or suspected bugs, which was itself quite valuable.
Happily, the Ledger project later revived and has attracted new active contributors. I have remained active in that community, sharing discoveries and design discussions, and we have seen many ideas travelling in both directions. hledger shared #ledger's IRC channel until 2014, when I added #hledger to allow us more space.
I think having independent but compatible implementations has been quite helpful for troubleshooting, exploring the design space, and growing the "Ledger-likes" community. My other projects in that direction include the ledger-cli.org site, LedgerTips, IRC support on #ledger, and now plaintextaccounting.org.
Compared to Ledger, hledger builds quickly and has a complete and accurate manual, an easier report query syntax, multi-column balance reports, better depth limiting, an interactive data entry assistant, and optional web and curses interfaces.
Compared to hledger, Ledger has additional power-user features such as periodic and modifier transactions, budget reports, and the built in value expressions language, and it remains faster and more memory efficient (for now).
We currently support:
We do not support:
And we add these commands:
hledger's journal file format is mostly identical with Ledger's, by design. Generally, it's easy to keep a journal file that works with both hledger and Ledger if you avoid Ledger's and hledger's more specialised syntax (or keep it in separate files which you include only when appropriate).
Some Ledger syntax is parsed but ignored (such as automated transactions and periodic transactions). Some features are not currently parsed and will cause an error, eg Ledger's more recent top-level directives. There can also be subtle differences in parser behaviour, such as with hledger comments vs Ledger comments, or balance assertions.
hledger recognises description and negative patterns by "desc:" and "not:" prefixes, unlike Ledger 3's free-form parser
hledger does not require a space between command-line flags and their values, eg -fFILE works as well as -f FILE
hledger's weekly reporting intervals always start on mondays
hledger shows start and end dates of the intervals requested, not just the span containing data
hledger always shows time balances (from the timeclock/timedot formats) in hours, with two decimal places
hledger splits multi-day time sessions at midnight by default (Ledger does this with an option)
hledger's output follows the decimal point character, digit grouping, and digit group separator character used in the journal.
hledger print shows amounts for all postings, and shows unit prices for amounts which have them. (This means that it does not currently print multi-commodity transactions in valid journal format.)
hledger print ignores the --date2 flag, always showing both dates. ledger print shows only the secondary date with --aux-date, but not vice versa.
hledger's default commodity directive (D) sets the commodity to be used for subsequent commodityless amounts, and also sets that commodity's display settings if such an amount is the first seen. Ledger uses D only for commodity display settings and for the entry command.
hledger's include directive does not support shell glob patterns (eg include *.journal ), as Ledger's does.
when checking balance assertions hledger sorts the account's postings first by date and then (for postings with the same date) by parse order. Ledger checks assertions in parse order, ignoring dates.
Ledger allows amounts to have a fixed lot price (the {} syntax ?) and a regular price in any order (and uses whichever appears first). hledger requires the fixed lot price to come last (and ignores it).
hledger uses --ignore-assertions/-I to disable balance assertions. Ledger uses --permissive, and -I means something else (--prices).
hledger's -p option doesn't combine nicely with -b/-e/-D/-W/-M/-Q/-Y. Basically if there's a -p, all those others are ignored. There's an open issue. With hledger you can also specify start and/or end dates with a query argument, date:START-END, which probably doesn't combine perfectly with the options.
in hledger version 1.3 onward, the "uncleared" status has been renamed to "unmarked", it is matched by the -U/--unmarked flag. Also, the --unmarked/--pending/--cleared flags can be combined, so eg -UP matches unmarked and pending, similar to Ledger's --uncleared flag. (#564)
hledger's -P flag is short for --pending. Ledger uses it for grouping by payee.
There is a ledger4 repo on github; this is John's 2012/2013 rewrite of some parts of Ledger 3, including the parser, in Haskell. We have a plan to add this parser to hledger in 2015/2016, increasing its ability to read Ledger's files.
This is an easy mistake at first. This journal entry:
1/1
a 1
bwill give a parse error (...can't have more than one real posting with no amount...).
There must always be at least two spaces between the account name and amount. So instead, it should be:
1/1
a 1
bWhen hledger needs to show a multi-commodity amount, each commodity is displayed on its own line, one above the other (like Ledger).
Here are some examples. With this journal, the implicit balancing amount drawn from the b account will be a multicommodity amount (a euro and a dollar):
2015/1/1
a EUR 1
a USD 1
bthe print command shows the b posting's amount on two lines, bottom-aligned:
$ hledger -f t.j print
2015/01/01
a USD 1
a EUR 1
EUR -1 ; <-
b USD -1 ; <- a euro and a dollar is drawn from bthe balance command shows that both a and b have a multi-commodity balance (again, bottom-aligned):
$ hledger -f t.j balance
EUR 1 ; <-
USD 1 a ; <- a's balance is a euro and a dollar
EUR -1 ; <-
USD -1 b ; <- b's balance is a negative euro and dollar
--------------------
0while the register command shows (top-aligned, this time!) a multi-commodity running total after the second posting, and a multi-commodity amount in the third posting:
$ hledger -f t.j register --width 50
2015/01/01 a EUR 1 EUR 1
a USD 1 EUR 1 ; <- the running total is now a euro and a dollar
USD 1 ;
b EUR -1 ; <- the amount posted to b is a negative euro and dollar
USD -1 0 ;Newer reports like multi-column balance reports show multi-commodity amounts on one line instead, comma-separated. Although wider, this seems clearer and we should probably use it more:
$ hledger -f t.j balance --yearly
Balance changes in 2015:
|| 2015
===++================
a || EUR 1, USD 1
b || EUR -1, USD -1
---++----------------
|| 0 You will also see amounts without a corresponding account name if you remove too many account name segments with --drop:
$ hledger -f t.j balance --drop 1
EUR 1
USD 1
EUR -1
USD -1
--------------------
0