Basic analysis of ledger files

ledger.R

#!/usr/bin/Rscript
printf <- function(fmt, ...) cat(sprintf(fmt, ...))

# Reads a Ledger file by doing CSV export on the CLI, and then slurping
# it into R's CSV reader.
#
# start and end can be used to limit which entries are allowed in by time,
# and accounts can be used to limit which accounts are read.
ledger.read <- function(filename,
                        start=NA,
                        end=NA,                        
                        accounts=NULL) {

    args <- c('-f', filename)
    if (!is.na(start)) args <- append(args, c('-b', start))
    if (!is.na(end)) args <- append(args, c('-e', end))
    args <- append(args, c('csv', accounts))

    raw <- system2('ledger', args, stdout=TRUE)

    conn <- textConnection(raw)
    data <- read.csv(
            conn,
            header=FALSE,
            col.names=c(
                "date", "tag", "description", "account", "currency",
                "amount", "X1", "X2"),
            stringsAsFactors=FALSE)

    close(conn)
    data
}

external <- ledger.read('accounts.ledger')

paydate <- NA
income <- 0
spent <- 0
transfers <- list()

# This is designed to track where the money from each paycheck
# ends up, by recording all account credits that occur since
# each payday. A couple of assumptions on the structure 
# of the accounts:
#
# - All work income is in a subaccount of Income:Work:
# - All expenses are in a subaccount of Expenses:
# - Virtual transactions (parenthesized) are ignored

invisible(apply(external,
      1,
      function (txn) {
        now <<- txn[[1]]
        account <- txn[[4]]
        amount <- as.numeric(txn[[6]])

        if (grepl('^Income:Work:', account)) {
            if (!is.na(paydate)) {
                printf('%s\n', strrep('-', 20))
                printf("Between %s And %s\n", paydate, now)
                printf("  Income: $%f\n", income)
                printf("  Spent:  $%f\n\n", spent)
                print(t(data.frame(transfers)))
                printf("\n\n")
            }

            income <<- -amount
            spent <<- 0
            transfers <<- list()
            paydate <<- now
        } else if (!grepl('^\\(', account)) {
            if (grepl('^Expenses:', account)) {
                spent <<- spent + amount
            } 

            # Not having something like this would show every transaction 
            # twice (credits and debits), including transfers to accounts
            # which are less interesting like Income
            if (amount > 0) {
                if (is.null(transfers[[account]])) {
                    transfers[[account]] <<- 0
                }

                transfers[[account]] <<- transfers[[account]] + amount
            }
        }
      }))

if (!is.na(paydate)) {
        printf('%s\n', strrep('-', 20))
        printf("Since %s\n", paydate)
        printf("  Income: $%f\n", income)
        printf("  Spent:  $%f\n\n", spent)
        print(t(data.frame(transfers)))
        printf("\n\n")
}