Joda-Time: formatting with multiple formats with one Formatter

Question

I would like to print durations given in milliseconds with different format specifications depending on its size:

case (1)      "H:mm"  if duration < 10 hours
case (2)     "HH:mm"  if duration < 24 hours
case (3)  "#d HH:mm"  else (duration >= 24 hours)

which means only a 1-hour field digit for less than 10 hours, but 2-hour field digits if there is an initial daily field!

Examples:

case (1)      "0:45"  means 45 minutes,
              "1:23"  means 1 hour and 23 minutes,
case (2)     "12:05"  means 12 hours and 5 minutes and
case (3)  "1d 05:09"  means 1 day, 5 hours and 9 minutes
                               (= 29 hours and 9 minutes).

I tried with

object JodaTest {
  import org.joda.time._
  private val pdf = {
    import format._
    val pfb = new PeriodFormatterBuilder()
      .appendDays.appendSeparator("d ")
      .printZeroAlways
      .minimumPrintedDigits(2).appendHours.appendSeparator(":")
      .appendMinutes
    new PeriodFormatter(pfb.toPrinter, null)
  }
  def durstr(duration: Long): String =
    pdf.print((new Period(duration)).normalizedStandard)
}

that leads to

  2700000 => "00:45"     but should be "0:45"
  4980000 => "01:23"     but should be "1:23"
 43500000 => "12:05"
104940000 => "1d 05:09"

but I don’t know how to omit the initial zero of the two-digit representation of the day in case (1), but at the same time I am forced to print it in case (3) with the same period.

Is it possible to do this with one org.joda.time.format.PeriodFormatter?

Answer 1

PeriodPrinter , , , .

Answer 2

, , , PeriodFormatter,

object JodaTest {
  import org.joda.time._
  import format._
  private def pdf(digits: Int) = new PeriodFormatter(
    new PeriodFormatterBuilder()
      .appendDays.appendSeparator("d ")
      .printZeroAlways
      .minimumPrintedDigits(digits).appendHours.appendSeparator(":")
      .minimumPrintedDigits(2).appendMinutes
      .toPrinter, null)
  private lazy val pdf1 = pdf(1)
  private lazy val pdf2 = pdf(2)
  def durstr(duration: Long): String = {
    val period = new Period(duration).normalizedStandard
    val pdf = if (period.getDays > 0) pdf2 else pdf1
    pdf.print(period)
  }
}

  2700000 => "0:45"
  4980000 => "1:23"
 43500000 => "12:05"
104940000 => "1d 05:09".

Answer 3

- PeriodFormatter, Joda-Time.

• , ​​ "000d 00:00" Joda-Time

object JodaTest {
  import org.joda.time._
  import format._
  // "000d 00:00" - 3 day digits for periods with up to 999 days long
  private val pdf = new PeriodFormatter(new PeriodFormatterBuilder()
    .printZeroAlways
    .minimumPrintedDigits(3).appendDays.appendSeparator("d ")
    .minimumPrintedDigits(2).appendHours.appendSeparator(":").appendMinutes
    .toPrinter, null)
  private def adjust(rawstr: String): String = {
    // "000d 00:00" => ("000d 0", "0:00")
    val (first, second) = rawstr splitAt 6
    // remove unwanted leading zeros in first part, keep it in second
    first.dropWhile(c => !c.isDigit || c == '0') + second
  }
  def durstr(duration: Long): String = {
    // PeriodType.dayTime => day is the most significant field (no weeks etc.)
    adjust(pdf.print(new Period(duration) normalizedStandard PeriodType.dayTime))
  }
}

   duration =>       rawstr =>       adjust
          0 => "000d 00:00" =>       "0:00"
    2700000 => "000d 00:45" =>       "0:45"
    4980000 => "000d 01:23" =>       "1:23"
   43500000 => "000d 12:05" =>      "12:05"
  104940000 => "001d 05:09" =>   "1d 05:09"
  518760000 => "006d 00:06" =>   "6d 00:06"
  605220000 => "007d 00:07" =>   "7d 00:07"
  951060000 => "011d 00:11" =>  "11d 00:11"
43230000000 => "500d 08:20" => "500d 08:20"

Of course, it would be nice to create such formats directly from Joda-Time by specifying a template, such as the Excel number format (#, ## 0.00 #), to say where zeros are needed ever or only when necessary. But it’s unclear how to define it precisely because you have not only “0” and “#”, but you also need characters for each field and putting literals in the format string (possibly through escaping) would also be nice.