User Tools

Site Tools

User Manual

This manual is for hledger 0.22.98 (the latest pre-0.23 HEAD).

It is partly an introduction, but mainly a reference manual for hledger. For a more gradual introduction, see step-by-step and more docs.


hledger is a program for tracking money, time, or any other commodity, using a simple, editable file format and double-entry accounting, inspired by and largely compatible with ledger. hledger is Free Software released under GPL version 3 or later.

hledger’s basic function is to read a plain text file describing (eg) financial transactions, and quickly generate useful reports via the command line. It can also help you record transactions, or (via add-ons) provide a local web interface for editing, or publish live financial data on the web.

You can use it to, eg:

  • track spending and income
  • track unpaid or due invoices
  • track time and report by day/week/month/project
  • get accurate numbers for client billing and tax filing

hledger works on linux, mac and windows. People most often build the latest release with cabal-install, like so:

$ cabal update
$ cabal install hledger [hledger-web]
$ hledger --version
hledger 0.22.1

For more help with this, and other install options, see the Installation Guide.

Basic Usage

Basic usage is:


Most commands query or operate on a journal file, which by default is .hledger.journal in your home directory. You can specify a different file with the -f option or LEDGER_FILE environment variable, or standard input with -f-.

Options are similar across most commands, with some variations; use hledger COMMAND --help for details. Most options must appear somewhere after COMMAND, not before it. These input and help-related options can appear anywhere: -f, --rules-file, --alias, --help, --debug, --version.

Arguments are also command-specific, but usually they form a query which selects a subset of the journal, eg transactions in a certain account.

To create an initial journal, run hledger add and follow the prompts to enter some transactions. Or, save this sample file as .hledger.journal in your home directory. Now try commands like these:

$ hledger                               # show available commands
$ hledger add                           # add more transactions to the journal file
$ hledger balance                       # all accounts with aggregated balances
$ hledger balance --help                # show help for balance command
$ hledger balance --depth 1             # only top-level accounts
$ hledger register                      # show a register of postings from all transactions
$ hledger reg income                    # show postings to/from income accounts
$ hledger reg checking                  # show postings to/from checking account
$ hledger reg desc:shop                 # show postings with shop in the description
$ hledger activity                      # show transactions per day as a bar chart

Data format

Journal files

hledger’s usual data source is a plain text file containing journal entries in hledger journal format. This file represents a standard accounting general journal. I use file names ending in .journal, but that’s not required. The journal file contains a number of transaction entries, each describing a transfer of money (or any commodity) between two or more named accounts, in a simple format readable by both hledger and humans.

hledger’s journal format is a compatible subset, mostly, of ledger's journal format, so hledger can work with compatible ledger journal files as well. It’s safe, and encouraged, to run both hledger and ledger on the same journal file, eg to validate the results you’re getting.

You can use hledger without learning any more about this file; just use the add or web commands to create and update it. Many users, though, also edit the journal file directly with a text editor, perhaps assisted by the helper modes for emacs or vim.

Here’s an example:

; A sample journal file. This is a comment.

2008/01/01 income               ; <- transaction's first line starts in column 0, contains date and description
    assets:bank:checking  $1    ; <- posting lines start with whitespace, each contains an account name
    income:salary        $-1    ;    followed by at least two spaces and an amount

2008/06/01 gift
    assets:bank:checking  $1    ; <- at least two postings in a transaction
    income:gifts         $-1    ; <- their amounts must balance to 0

2008/06/02 save
    assets:bank:saving    $1
    assets:bank:checking        ; <- one amount may be omitted; here $-1 is inferred

2008/06/03 eat & shop           ; <- description can be anything
    expenses:food         $1
    expenses:supplies     $1    ; <- this transaction debits two expense accounts
    assets:cash                 ; <- $-2 inferred

2008/12/31 * pay off            ; <- an optional * or ! after the date means "cleared" (or anything you want)
    liabilities:debts     $1

Now let’s explore the available journal file syntax in detail.


Each journal entry begins with a simple date in column 0, followed by three optional fields with spaces between them: a status flag (* or ! or nothing), a transaction code (eg a check number), and/or a description; then two or more postings (of some amount to some account), each on their own line.

The posting amounts within a transaction must always balance, ie add up to 0. You can leave one amount blank and it will be inferred.


Simple dates

Within a journal file, transaction dates always follow a year/month/day format, although several different separator characters are accepted. Some examples: 2010/01/31, 2010/1/31, 2010-1-31, 2010.1.31.

Writing the year is optional if you set a default year with a Y directive. This is a line containing Y and the year; it affects subsequent transactions, like so:


12/15  ; equivalent to 2009/12/15


1/31  ; equivalent to 2010/1/31
Secondary dates

Real-life transactions sometimes involve more than one date - eg the date you write a cheque, and the date it clears in your bank. When you want to model this, eg for more accurate balances, write both dates separated by an equals sign. The primary date, on the left, is used by default; the secondary date, on the right, is used when the --date2 flag is specified (--aux-date or --effective will also work).

Their meaning is up to you, but it’s best to follow a consistent rule. I write the bank’s clearing date as primary, and the date I initiated the transaction as secondary (if needed).


; The secondary date's year is optional, defaulting to the primary's
2010/2/23=2/19 movie ticket
  expenses:cinema                   $10

$ hledger register checking
2010/02/23 movie ticket         assets:checking                $-10         $-10

$ hledger register checking --date2
2010/02/19 movie ticket         assets:checking                $-10         $-10
Posting dates

Comments and tags are covered below, but while we are talking about dates: you can give individual postings a different date from their parent transaction, by adding a posting tag like date:DATE, where DATE is a simple date. The secondary date can be set with date2:DATE2. If present, these dates will take precedence in reports.

Ledger’s bracketed posting date syntax ([DATE], [DATE=DATE2] or [=DATE2] in a posting comment) is also supported, as an alternate spelling of the date and date2 tags.

Note: if you do use either of these forms, be sure to give them a valid DATE or you’ll get a parse error, eg an empty date: tag is not allowed.


Account names typically have several parts separated by a full colon, from which hledger derives a hierarchical chart of accounts. They can be anything you like, but in finance there are traditionally five top-level accounts: assets, liabilities, income, expenses, and equity.

Account names may contain single spaces, eg: assets:accounts receivable.


After the account name, there is usually an amount. Important: between account name and amount, there must be two or more spaces.

The amount is a number, optionally with a currency symbol or commodity name on either the left or right. Negative amounts may have the minus sign either before or after the currency symbol (-$1 or $-1). Commodity names which contain more than just letters should be enclosed in double quotes (1 "person hours").

Decimal points and digit groups

hledger supports flexible decimal point and digit group separator styles, to support international variations. Numbers can use either a period (.) or a comma (,) as decimal point. They can also have digit group separators at any position (eg thousands separators) which can be comma or period - whichever one you did not use as a decimal point. If you use digit group separators, you must also include a decimal point in at least one number in the same commodity, so that hledger knows which character is which. Eg, write $1,000.00 or $1.000,00.

Canonical amount styles

Based on how you format amounts, hledger will infer canonical display styles for each commodity, and use these when displaying amounts in that commodity. Amount styles include:

  • the position (left or right) and spacing (space or no separator) of the commodity symbol
  • the digit group separator character (comma or period) and digit group sizes, if any
  • the decimal point character (period or comma)
  • the display precision (number of decimal places displayed)

The canonical style is generally the style of the first amount seen in a commodity (which may be in a default commodity directive. The precision is the highest precision seen among all amounts in the commmodity.

Balance Assertions

hledger supports ledger-style balance assertions in journal files. These look like =EXPECTEDBALANCE following a posting’s amount. Eg in this example we assert the expected balance in accounts a and b after each posting:

  a   $1  =$1
  b       =$-1

  a   $1  =$2
  b  $-1  =$-2

After reading a journal file, hledger will check all balance assertions and report an error if any of them fail. Balance assertions can protect you from, eg, inadvertently disrupting reconciled balances while cleaning up old entries.

Note, when checking balance assertions hledger sorts the account’s postings first by date and then (for postings with the same date) by parse order. This is different from ledger, which currently goes strictly by parse order. Sorting by date means balance assertions will still work if you reorder your entries.

With included files, things are a little more complicated. Including preserves the ordering of postings and assertions. If you have multiple postings to an account on the same day, split across different files, and you also want to assert the account’s balance on the same day, you can run into trouble depending on where the assertion is located.

Also note the asserted balance must be a simple amount - it’s not currently possible to assert a balance containing multiple commodities.

The impact of many balance assertions on parsing time for large files is unknown.


When recording an amount, you can also record its price in another commodity. This documents an exchange rate that was applied within this transaction (or to be precise, within the posting). There are three ways to specify a transaction price:

  1. Write the unit price (exchange rate) explicitly as @ UNITPRICE after the amount:

    2009/1/1 assets:foreign currency €100 @ $1.35 ; one hundred euros at $1.35 each assets:cash

  2. Or write the total price for this amount as @@ TOTALPRICE:

    2009/1/1 assets:foreign currency €100 @@ $135 ; one hundred euros at $135 for the lot assets:cash

  3. Or fully specify all posting amounts using exactly two commodities:

    2009/1/1 assets:foreign currency €100 ; one hundred euros assets:cash $-135 ; exchanged for $135

You can use the --cost/-B flag with reporting commands to see such amounts converted to their price’s commodity. Eg, using any of the above examples we get:

$ hledger print --cost
    assets:foreign currency       $135.00
    assets                       $-135.00
Fixed Lot Prices

ledger has another syntax for fixed lot prices. ({=PRICE}). In ledger, this is equivalent to @ PRICE, except you can provide both and then ledger generates an automatic Capital Losses posting covering the difference.

hledger will parse this syntax, but ignore it.

Historical prices

hledger will parse and ignore ledger-style historical price directives:

    ; Historical price directives look like: P DATE COMMODITYSYMBOL UNITPRICE
    ; These say the euro's exchange rate is $1.35 during 2009 and
    ; $1.40 from 2010/1/1 on.
    P 2009/1/1 € $1.35  
    P 2010/1/1 € $1.40


  • A semicolon (;) or hash (#) in column 0 starts a journal comment line, which hledger will ignore.
  • A semicolon after a transaction’s description and/or indented on the following lines starts a transaction comment.
  • A semicolon after a posting’s amount and/or indented on the following lines starts a posting comment.

Transaction and posting comments are displayed by print, can contain tags and can be queried.

Some examples:

# a journal comment

; also a journal comment
2012/5/14 something  ; a transaction comment
    ; the transaction comment, continued
    posting1  1  ; a comment for posting 1
    ; a comment for posting 2
    ; another comment line for posting 2
; a journal comment (because not indented)


You can include tags (labels), optionally with values, in transaction and posting comments, and then query by tag. This is like Ledger’s metadata feature, except hledger’s tag values are simple strings.

A tag is any unspaced word immediately followed by a full colon, eg: sometag: . A tag’s value is the text following the colon, if any, until the next newline or comma, with leading and trailing whitespace removed. Comma may be used to write multiple tags on one line.

For example, here is a transaction with three tags, the posting has one, and all tags have values except TAG1:

1/1 a transaction    ; TAG1:, TAG2: tag2's value
    ; TAG3: a third transaction tag
    a  $1  ; TAG4: a posting tag


Account aliases

You can define account aliases to rewrite certain account names (and their subaccounts). This tends to be a little more reliable than post-processing with sed or similar. The directive is alias ORIG = ALIAS, where ORIG and ALIAS are full account names. Eg:

alias expenses = equity:draw:personal

To forget all aliases defined to this point, use:

end aliases

You can also specify aliases on the command line:

$ hledger --alias 'my earning=income:business' ...

Journal directive aliases are applied first, then command-line aliases, and at most one of each will be applied to each account name.

See also How to use account aliases.

Default commodity

You can set a default commodity, to be used for any subsequent amounts which have no commodity symbol (including added amounts), with the D directive:

; set british pound as default commodity
; also sets canonical style for pound amounts, since it's the first one
D £1,000.00

  a  2340    ; no symbol, will use default commodity and style
             ; (pound symbol on left, comma thousands separator, two decimal places) 

A default commodity directive can also influence the canonical amount style for the commodity, as in the example above.

Default parent account

You can specify a parent account which will be prepended to all accounts within a section of the journal. Use the account directive like so:

account home

    food    $10


If !end is omitted, the effect lasts to the end of the file. The above is equivalent to:

    home:food           $10
    home:cash          $-10

Included files are also affected, eg:

account business
include biz.journal
account personal
include personal.journal
Including other files

You can pull in the content of additional journal files, by writing lines like this:

include path/to/file.journal

The include directive may only be used in journal files, and currently it may only include other journal files (eg, not CSV or timelog files.)

CSV files

hledger can also read CSV files, translating the CSV records into journal entries on the fly. We must provide some some conversion hints in a “rules file”, named like the CSV file with an extra .rules suffix (you can choose another name with --rules-file).
If the rules file does not exist, it will be created with default rules, which you’ll need to tweak. Here’s a minimal rules file. It says that the first and second CSV fields are the journal entry’s date and amount:

fields date, amount

Lines beginning with # or ; and blank lines are ignored. The following kinds of rule can appear in any order:

(Field list) This names the CSV fields (names may not contain whitespace or ; or #), and also assigns them to journal entry fields when you use any of these names:


(Field assignment) This assigns the given text value to a journal entry field (one of the field names above). CSV fields can be referenced with %CSVFIELDNAME or %N (N starts at 1) and will be interpolated.

You can use a field list, field assignments, or both. At least the date and amount fields must be assigned.

(Conditional block) This applies the field assignments only to CSV records matched by one of the PATTERNS.

PATTERNS is one or more regular expressions, each on its own line. The first pattern can optionally be written on the same line as the if; patterns on the following lines must start in column 0 (no indenting). The regular expressions are case insensitive, and can match anywhere within the whole CSV record. (It’s not yet possible to match within a specific field.)

FIELDASSIGNMENTS is one or more field assignments (described above), each on its own line and indented by at least one space. (The indent is required for successful parsing.)

Example 1. The simplest conditional block has a single pattern and a single field assignment. Here, any CSV record containing the pattern groceries will have its account2 field set to expenses:groceries.

if groceries
 account2 expenses:groceries

Example 2. Here, CSV records containing any of these patterns will have their account2 and comment fields set as shown. The capitalisation is not required, that’s just how I copied them from my bank’s CSV.

 account2 expenses:business:banking
 comment  XXX probably deductible, check

skip [N]
Skip this number of CSV records (1 by default). Use this to skip CSV header lines.

date-format DATEFMT
This is required if the values for date or date2 fields are not in YYYY/MM/DD format (or close to it). DATEFMT specifies a strptime-style date parsing pattern containing year/month/date format codes. Note the pattern must parse the CSV date value completely. Some examples:

# "6/11/2013"
date-format %-d/%-m/%Y
# "11/06/2013"
date-format %m/%d/%Y
# "2013-Nov-06"
date-format %Y-%h-%d
# "11/6/2013 11:32 PM"
date-format %-m/%-d/%Y %l:%M %p

Include another rules file at this point. Useful for common rules shared across multiple CSV files.

Typically you’ll keep one rules file for each account which you download as CSV. For an example, see How to read CSV files.

Other notes:

An amount value that is parenthesised will have the parentheses stripped and its sign flipped.

If the currency pseudo field is assigned, its value will be prepended to every amount.

If the CSV has debit/credit amounts in separate fields, assign the amount-in and amount-out pseudo fields instead of amount.

Generating entries with three or more postings is not supported at present.

Timelog files

hledger can also read time log files. These are (a subset of) timeclock.el’s format, containing clock-in and clock-out entries like so:

i 2009/03/31 22:21:45 projects:A
o 2009/04/01 02:00:34

hledger treats the clock-in description (“projects:A”) as an account name, and creates a virtual transaction (or several - one per day) with the appropriate amount of hours. From the time log above, hledger print gives:

2009/03/31 * 22:21-23:59
    (projects:A)          1.6h

2009/04/01 * 00:00-02:00
    (projects:A)          2.0h

Here is a sample.timelog to download and some queries to try:

hledger -f sample.timelog balance                               # current time balances
hledger -f sample.timelog register -p 2009/3                    # sessions in march 2009
hledger -f sample.timelog register -p weekly --depth 1 --empty  # time summary by week

To generate time logs, ie to clock in and clock out, you could:

  • use emacs and the built-in timeclock.el, or the extended timeclock-x.el and perhaps the extras in ledgerutils.el
  • at the command line, use these bash aliases: alias ti="echo i `date '+%Y-%m-%d %H:%M:%S'` \$* >>$TIMELOG" alias to="echo o `date '+%Y-%m-%d %H:%M:%S'` >>$TIMELOG"
  • or use the old ti and to scripts in the ledger 2.x repository. These rely on a “timeclock” executable which I think is just the ledger 2 executable renamed.


hledger provides a number of subcommands out of the box; run hledger with no arguments to see a list. More add-on commands will appear if you install additional hledger-* packages, or if you put programs or scripts named hledger-NAME in your PATH.

To run a command, you just need to type its unique prefix, eg hledger reg is a shortcut for hledger register. (Also, hledger bs is short for hledger balancesheet.)

Data entry

Many hledger users edit their journals directly with a text editor, or generate them from CSV. For more interactive data entry, there is the add command and also the web add-on (below).


The add command prompts interactively for new transactions, and appends them to the journal file. Just run hledger add and follow the prompts. You can add as many transactions as you like; when you are finished, press control-d or control-c to exit.

Additional convenience features:

  • Sensible defaults are provided where possible. You can set the initial defaults by providing them as command line arguments. If there is a recent transaction with a description similar to the one you entered, it will be displayed and used for defaults.
  • Readline-style edit keys may be used during data entry. Eg control-p recalls previous entries.
  • While entering account names, the tab key will auto-complete or list the available completions, based on the existing transactions.
  • If the journal defines a default commodity, it will be added to any bare numbers entered.
  • A code (in parentheses) may be entered at the Date: prompt, following the date. Comments and/or tags may be entered following a date or amount.
  • If you make a mistake, enter < at any prompt to restart the transaction.

An example:

$ hledger add
Starting a new transaction.
date ? [2013/04/09]: 
description ? : starbucks

Using this existing transaction for defaults:
2012/04/19 * starbucks
    expenses:personal:food:snacks         $3.70
    assets:cash:wallet                   $-3.70

account 1 ? [expenses:personal:food:snacks]: 
amount  1 ? [$3.7]: 
account 2 ? [assets:cash:wallet]: 
amount  2 ? [$-3.7]: 
account 3 (or . to complete this transaction) ? : .

Transaction entered:
2013/04/09 starbucks
    expenses:personal:food:snacks          $7.7
    assets:cash:wallet                    $-7.7

Accept this transaction ? [y]: 
Added to the journal.

Starting a new transaction.
date ? [2013/04/09]: <CTRL-D>


These are the commands for actually querying your ledger. The most basic reporting commands are print, register and balance:


The print command displays full transactions from the journal file, tidily formatted and showing all amounts explicitly. The output of print is always a valid hledger journal, but it does always not preserve all original content exactly (eg directives).

hledger’s print command also shows all unit prices in effect, or (with -B/–cost) shows cost amounts.


$ hledger print
$ hledger print employees:bob | hledger -f- register expenses


The register command displays postings, one per line, and their running total. With no query terms, this is not all that different from print:

$ hledger register

More typically, use it to see a specific account’s activity:

$ hledger register assets:bank:checking

The --depth option limits the amount of sub-account detail displayed:

$ hledger register assets:bank:checking --depth 2

With a reporting interval it shows aggregated summary postings within each interval:

$ hledger register --monthly rent
$ hledger register --monthly -E food --depth 4

The --average/-A flag shows the running average posting amount instead of the running total.

The --related/-r flag shows the other postings in the transactions of the postings which would normally be shown.

The --width/-w option adjusts the width of the output. By default, this is 80 characters. To allow more space for descriptions and account names, use -w to increase the width to 120 characters, or -wN to set any desired width (at least 50 recommended). Note, currently -w/–width can not have a space between flag and value (#149).


The balance command displays accounts and their balances, indented to show the account hierarchy. Examples:

$ hledger balance
$ hledger balance food -p 'last month'

Accounts which have zero balance (and no non-zero subaccounts) will be omitted by default; use -E/--empty to show them. “Boring parent” accounts, which contain a single interesting subaccount and no balance of their own, are elided into the subaccount’s line for more compact output; use --no-elide to prevent this.

A final total is displayed, use --no-total to suppress this. Also, the --depth N option shows accounts only to the specified depth, useful for an overview:

$ for y in 2006 2007 2008 2009 2010; do echo; echo $y; hledger -f $y.journal balance ^expenses --depth 2; done

With --flat, a simple list of full account names is displayed instead, each with it’s inclusive balance (including any subaccount balances). In this mode you can use --drop N to elide the first few account name components. The --depth flag also works.

With a reporting interval, multiple columns will be shown, one for each period. There are three modes available:

  1. By default each column shows the sum of postings in that period, ie the account’s change of balance in that period. This is good for a multi-column income statement:

    $ hledger balance ^income ^expense -p ‘monthly this year’ –depth 3

    or cashflow statement:

    $ hledger balance ^assets ^liabilities ‘not:(receivable|payable)' -p ‘weekly this month’

  2. With --cumulative, the report shows the ending balance for each account at the end of each period, starting from zero at the report start date.
  3. With --historical/-H, it shows the actual ending balance at the end of each period, including any balance from postings before the report start date. This is good for historical balance sheets:

    $ hledger balance ^assets ^liabilities -YH


This command displays a simple income statement. It currently assumes that you have top-level accounts named income (or revenue) and expense (plural forms also allowed.)


This command displays a simple balance sheet. It currently assumes that you have top-level accounts named asset and liability (plural forms also allowed.)


This command displays a simplified cashflow statement (without the traditional segmentation into operating, investing, and financing cash flows.) It shows the change in all “cash” accounts for the period. It currently assumes that cash accounts are under a top-level account named asset and do not contain receivable or A/R (plural forms also allowed.)


The activity command displays a simplistic textual bar chart showing transaction counts by day, week, month or other reporting interval.


$ hledger activity -p weekly dining


The stats command displays summary information for the whole journal, or a matched part of it.


$ hledger stats
$ hledger stats -p 'monthly in 2009'



This command runs hledger’s built-in unit tests and displays a quick report. A pattern can be provided to filter tests by name. It’s mainly used in development, but it’s also nice to be able to check hledger for smoke at any time.


$ hledger test
$ hledger test -v balance


Add-on commands are executables named hledger-* installed in your PATH. hledger will detect these at startup and offer them as extra commands. Run hledger without a command to see a list.

hledger-web is released along with hledger and supported on all the major platforms, while other add-ons may or may not be.


ledger-autosync, which includes a hledger-autosync alias, downloads transactions from your bank(s) via OFX, and prints just the new ones as journal entries which you can add to your journal. It can also operate on .OFX files which you’ve downloaded manually. It can be a nice alternative to hledger’s built-in CSV reader, especially if your bank supports OFX download.


hledger-interest computes interests for a given account. Using command line flags, the program can be configured to use various schemes for day-counting, such as act/act, 30/360, 30E/360, and 30/360isda. Furthermore, it supports a (small) number of interest schemes, i.e. annual interest with a fixed rate and the scheme mandated by the German BGB288 (Basiszins für Verbrauchergeschäfte). See the package page for more.


hledger-irr computes the internal rate of return, also known as the effective interest rate, of a given investment. After specifying what account holds the investment, and what account stores the gains (or losses, or fees, or cost), it calculates the hypothetical annual rate of fixed rate investment that would have provided the exact same cash flow. See the package page for more.


hledger-web provides a web-based user interface for viewing and modifying your ledger (demo). It includes an account register view that is more useful than the command-line register, and basic data entry and editing.

web-specific options:

--server            log requests, don't exit on inactivity
--port=N            serve on tcp port N (default 5000)
--base-url=URL      use this base url (default http://localhost:PORT/)
--static-root=URL   use this base url for static files (default http://localhost:PORT/static)

By default, the web command starts a transient local web app and displays it in your default web browser (“local ui mode”). With --server, it starts the web app, leaves it running, and also logs requests to the console (“server mode”).

Typically in server mode you’ll also want to use --base-url to set the protocol/hostname/port/path to be used in hyperlinks.

You can use --port to listen on a different TCP port, eg if you are running multiple hledger-web instances. Note --port‘s argument need not be the same as the PORT in the base url.

The more advanced option --static-root allows the static files served from a separate base url. This enables the optimization that the static files can be served from a generic web server like apache, which is good at handling static files and caching. One can also serve the files in a separate domain to reduce cookies overhead.

Note: unlike any other hledger command, web can alter existing journal data, via the edit form. A numbered backup of the file is saved on each edit, normally (ie if file permissions allow, disk is not full, etc.) Also, there is no built-in access control. So unless you run it behind an authenticating proxy, any visitor to your server will be able to see and overwrite the journal file (and included files.)

hledger-web disallows edits which would leave the journal file not in valid journal format. If the file becomes unparseable by other means, hledger-web will show an error until the file has been fixed.


$ hledger-web
$ hledger-web -E -B --depth 2 -f some.journal
$ hledger-web --server --port 5010 --base-url --debug

The following add-ons are examples and experiments provided in the extra directory in the hledger source. Add this directory to your PATH to make them available. The scripts are designed to run interpreted on unix systems (for tweaking), or you can compile them (for speed and robustness).


Prints all account names in the default journal.


Like the balance command, but with CSV output.


Like ledger’s equity command, this prints a single journal entry with postings matching the current balance in each account (or the specified accounts) in the default journal. An entry like this is useful to carry over asset and liability balances when beginning a new journal file, eg at the start of the year.

You can also use the same entry with signs reversed to close out the old file, resetting balances to 0. This means you’ll see the correct asset/liability balances whether you use one file or a whole sequence of files as input to hledger.

Prints only journal entries which are unique (by description).


Like the register command, but with CSV output.

Common options

The following common features and options work with most subcommands.


Part of hledger’s usefulness is being able to report on just a precise subset of your data.
Most commands accept an optional query expression, written as arguments after the command name, to filter the data by date, account name or other criteria. Query expressions are also used in the web ui‘s search form.

The query syntax is similar to a Google search expression: one or more space-separated search terms, optional prefixes to match specific fields, quotes to enclose whitespace, etc. A query term can be any of the following:

  • REGEX - match account names by this regular expression
  • acct:REGEX - same as above
  • code:REGEX - match by transaction code (eg check number)
  • desc:REGEX - match transaction descriptions
  • date:PERIODEXPR - match dates within the specified period. Actually, full period syntax is not yet supported.
  • date2:PERIODEXPR - as above, but match secondary dates
  • tag:NAME[=REGEX] - match by (exact, case sensitive) tag name, and optionally match the tag value by regular expression. Note tag: will match a transaction if it or any its postings have the tag, and will match posting if it or its parent transaction has the tag.
  • depth:N - match (or display, depending on command) accounts at or above this depth
  • status:1 or status:0 - match cleared/uncleared transactions
  • real:1 or real:0 - match real/virtual-ness
  • empty:1 or empty:0 - match if amount is/is not zero
  • amt:N or amt:=N, amt:<N, amt:>N - match postings with a single-commodity amount equal to, less than, or greater than N. (Multi-commodity amounts are not tested, and always match.) This has two modes of operation: if N is preceded by a + or - sign, the actual signed numbers are compared; otherwise the unsigned absolute values are compared.
  • cur:REGEX - match postings or transactions including any amounts whose currency/commodity symbol is fully matched by REGEX. (For a partial match, use .*REGEX.*). Note, to match characters which are regex-significant, like the dollar sign ($), you need to prepend \. And when using the command line you need to add one more level of quoting to hide it from the shell, so eg do: hledger print cur:'\$' or hledger print cur:\\$.
  • not: before any of the above negates the match

How query terms combine

hledger query expressions don’t support full boolean logic. Instead, multiple query terms are combined as follows:

  • The print command selects transactions which:
    • match any of the description terms AND
    • have any postings matching any of the positive account terms AND
    • have no postings matching any of the negative account terms AND
    • match all the other terms.

  • Other reporting commands (eg register and balance) select transactions/postings/accounts which match (or negatively match):
    • any of the description terms AND
    • any of the account terms AND
    • all the other terms.

Query options vs query arguments

On the command line, some of the query terms above can also be expressed as command-line flags. Generally you can mix and match query arguments and flags, and the resulting query will be their intersection. Note within the command-line flags, a -p period flag causes any -b or -e flags, and any preceding -p flags, to be ignored.

Smart dates

Unlike the journal file format, hledger’s user interface accepts flexible “smart dates”, for example in the -b and -e options, period expressions, display expressions, the add command and the web add form. Smart dates allow some natural english words, will assume 1 where less-significant date parts are unspecified, and can be relative to today’s date. Examples:

  • 2009/1/1, 2009/01/01, 2009-1-1, 2009.1.1 (simple dates)
  • 2009/1, 2009 (these also mean january 1, 2009)
  • 1/1, january, jan, this year (relative dates, meaning january 1 of this year)
  • next year (january 1, next year)
  • this month (the 1st of the current month)
  • this week (the most recent monday)
  • last week (the monday of the week before this one)
  • today, yesterday, tomorrow

Spaces in smart dates are optional, so eg -b lastmonth or date:fromlastmonth are valid.

Period expressions

hledger supports flexible “period expressions” with the -p/--period option to select transactions within a period of time (eg in 2009) and/or with a reporting interval (eg weekly). hledger period expressions are similar but not identical to ledger’s.

Here is a basic period expression specifying the first quarter of 2009. Note the start date is always included and the end date is always excluded:

-p "from 2009/1/1 to 2009/4/1"

Keywords like “from” and “to” are optional, and so are the spaces. Just don’t run two dates together:

-p"2009/1/1 2009/4/1"

Dates are smart dates, so if the current year is 2009, the above can also be written as:

-p "1/1 to 4/1"
-p "january to apr"
-p "this year to 4/1"

If you specify only one date, the missing start or end date will be the earliest or latest transaction in your journal:

-p "from 2009/1/1"  (everything after january 1, 2009)
-p "from 2009/1"    (the same)
-p "from 2009"      (the same)
-p "to 2009"        (everything before january 1, 2009)

A single date with no “from” or “to” defines both the start and end date like so:

-p "2009"           (the year 2009;    equivalent to "2009/1/1 to 2010/1/1")
-p "2009/1"         (the month of jan; equivalent to "2009/1/1 to 2009/2/1")
-p "2009/1/1"       (just that day;    equivalent to "2009/1/1 to 2009/1/2")

The -b/--begin and -e/--end options may be used as a shorthand for -p 'from ...' and -p 'to ...' respectively.

Note, however: a -p/--period option in the command line will cause any -b/-e/-D/-W/-M/-Q/-Y flags to be ignored.

Reporting interval

Period expressions can also begin with (or be) a reporting interval, which affects commands like register and activity. The reporting interval can be daily, weekly, monthly, quarterly, yearly, or one of the every ... expressions below, optionally followed by in. Examples:

-p "weekly from 2009/1/1 to 2009/4/1"
-p "monthly in 2008"
-p "bimonthly from 2008"
-p "quarterly"
-p "every 2 weeks"
-p "every 5 days from 1/3"
-p "every 15th day of month"
-p "every 4th day of week"

A reporting interval may also be specified with the -D/--daily, -W/--weekly, -M/--monthly, -Q/--quarterly, and -Y/--yearly options. But as noted above, a -p/--period option will override these.

Display expressions

A period expression or other query selects the transactions to be used for calculation. A display expression, specified with -d/--display, selects a more limited subset of transactions to be displayed in the report output.

This useful, say, if you want to see your checking register just for this month, but with an accurate running balance based on all transactions. Eg:

$ hledger register checking --display "d>=[1]"

meaning “make a register report of all checking transactions, but display only the ones with date on or after the 1st of this month.” Any smart date can appear inside the brackets.

The above the only kind of display expression we currently support: transactions before or after a given date.

Depth limiting

With the --depth N option, reports will show only the uppermost accounts in the account tree, down to level N. See the balance, register and chart examples.

Custom output formats

The --format FMT option will customize the line format of the balance command’s output (only, for now). FMT is a C printf/strftime-style format string, with the exception that field names are enclosed in parentheses:


If the minus sign is given, the text is left justified. The MIN field specified a minimum number of characters in width. After the value is injected into the string, spaces is added to make sure the string is at least as long as MIN. Similary, the MAX field specifies the maximum number of characters. The string will be cut if the injected string is too long.

  • %-(total) the total of an account, left justified
  • %20(total) The same, right justified, at least 20 chars wide
  • %.20(total) The same, no more than 20 chars wide
  • %-.20(total) Left justified, maximum twenty chars wide

The following FIELD types are currently supported:

  • account inserts the account name
  • depth_spacer inserts a space for each level of an account’s depth. That is, if an account has two parents, this construct will insert two spaces. If a minimum width is specified, that much space is inserted for each level of depth. Thus %5_, for an account with four parents, will insert twenty spaces.
  • total inserts the total for the account


If you want the account before the total you can use this format:

$ hledger balance --format "%20(account) %-(total)"
              assets $-1
         bank:saving $1
                cash $-2
            expenses $2
                food $1
            supplies $1
              income $-2
               gifts $-1
              salary $-1
   liabilities:debts $1

Or, if you’d like to export the balance sheet:

$ hledger balance --format "%(total);%(account)" --no-total

The default output format is %20(total) %2(depth_spacer)%-(account)


Here are some issues you might encounter when you run hledger (and remember you can also seek help from the IRC channel, mail list or bug tracker):

Successfully installed, but "No command 'hledger' found"

cabal installs binaries into a special directory, which should be added to your PATH environment variable. On unix-like systems, it is ~/.cabal/bin.

Fails to parse some valid Ledger files

"Illegal byte sequence" or "Invalid or incomplete multibyte or wide character" errors

In order to handle non-ascii letters and symbols (like £), hledger needs an appropriate locale. This is usually configured system-wide; you can also configure it temporarily. The locale may need to be one that supports UTF-8, if you built hledger with GHC < 7.2 (or possibly always, I’m not sure yet).

Here’s an example of setting the locale temporarily, on ubuntu gnu/linux:

$ file my.journal
my.journal: UTF-8 Unicode text                 # <- the file is UTF8-encoded
$ locale -a
en_US.utf8                             # <- a UTF8-aware locale is available
$ LANG=en_US.utf8 hledger -f my.journal print   # <- use it for this command

Here’s one way to set it permanently, there are probably better ways:

$ echo "export LANG=en_US.UTF-8" >>~/.bash_profile
$ bash --login

If we preferred to use eg fr_FR.utf8, we might have to install that first:

$ apt-get install language-pack-fr
$ locale -a
$ LANG=fr_FR.utf8 hledger -f my.journal print

Note some platforms allow variant locale spellings, but not all (ubuntu accepts fr_FR.UTF8, mac osx requires exactly fr_FR.UTF-8).

manual.txt · Last modified: 2014/04/16 17:39 by · Currently locked by:,