Develop Reference

Text cannot be parsed to a Duration in java8

I am trying to parse a text to duration as follow:
final Duration duration = TimeUtil.parseDuration("1.0:00:00");
But I get the following error,
Text cannot be parsed to a Duration
so can anyone tells me where my problem is?

If you are using protocol buffers' TimeUtil you specify a duration with seconds separated from nanoseconds by a period. The value may be lead by a minus sign, if the duration is negative (so that adding the duration to a time would move into the past relative to the time). The string must end in "s".
You can see the pretty simple parse and toString of a protocol buffer's duration in the public git[hub] repo's TimeUtil.
Given the type of duration, I'm guessing they're used for calculations on date-times that are internally represented as signed 64 bit nano seconds since unix epoch.
In other words it looks like these are valid durations:
"1s" // one second forward
"1.0s" // one second forward
"1.01s" // one second, 10,000,000 nano seconds forward
"-1.01s" // one second, 10,000,000 nano seconds backward
"60s" // one minute forward
"-86400s" // yesterday (one day backward)
// [assuming no daylight saving changes or leap seconds happened]
The Protocol Buffers' TimeUtil.parseDuration would not give you the error message you say you got, and is not at all like Duration.parse, which is more clearly documented and might give that kind of error message.

If guessed API:
parseDuration(String duration) takes as parameters:
duration - "3h" or "2mn" or "7s" or null.
Taken from java.lang.Object ninja.utils.TimeUtil API. It returns the number of seconds.
Then "1.0:00:00" is obviously not to be parsed.

Time of the day in minutes Java

I am looking to calculate the number of minutes given the time of the day.
Eg.: when input is 11:34, the output should be 11*60+34. The date doesn't matter.
I only need it down to the minutes scale. Seconds, milliseconds... don't matter.
Is there a method somewhere in Java doing this the neat way without me calculating it?
Right now, i'm using theTime.split(":"), theTime is a String holding "11:34" here, parsing the integers on each side and doing the calculation.
I saw Time but what I'm doing right now seemed more direct.
Nothing in Systems either.

There is no build in method for it. However here is a one-liner for it:
int timeInMins = Calendar.getInstance().get(Calendar.HOUR_OF_DAY) * 60 + Calendar.getInstance().get(Calendar.MINUTE);

Your approach looks good and sound, however to answer your question it would be simple to say that there is no such build in method which does that. You have to calculate it the way you are doing it right now.

Hi maybe you could use JodaTime? Below example how to get number of minutes from parsed string and from current time. In java 8 there is similar api but I haven't found exactly method like minutesOfDay()
#Test
public void learnHowManyMinutesPassedToday() {
DateTime time = DateTimeFormat.forPattern("HH:mm").parseDateTime("11:34");
System.out.println(time.getMinuteOfDay());
System.out.println(DateTime.now().getMinuteOfDay());
}

If you are looking to have input not from a String, take a look at
Java.util.Calendar.
It has Calendar.HOUR_OF_DAY and Calendar.HOUR and Calendar.MINUTE which could be your input. I'm not sure what the "neat" way of doing this would be. It is a simple calculation.

Calendar rightNow = Calendar.getInstance();
int hour = rightNow.get(Calendar.HOUR);
int min = rightNow.get(Calendar.MINUTE);
System.out.println("TimeMinutes:" + hour * 60 + min);
EDIT:
Except using split use the above.

java regex match up to a string

Regex'ers:
How can I construct a Java Regex to match Strings lexigraphically <= to a given date string?
For example, suppose the input is in YYYY-DD-MM format:
2014-01-20 MLK day
2007-04-14 'twas a very good day
2014-05-19 is today
1998-11-30 someone's birthday
I'd like the filter to return all lines before, say, Groundhog's day of this year, 2014-02-20;
so in the above list the regex would return all lines except today. (I don't want to convert the
dates to Epoch time; I'd like to just pass a Regex to a class that runs a map/reduce job so that
my input record reader can use the Regex as it constructs bundles to deliver to the mappers.)
TIA,

It's near impossible to do <= type logic with regular expressions. You technically could, but you'd have to map out every possible scenario...and then if you want to change the date you are comparing to, the whole expression would change. Instead, I'd just match all the dates/values and then use a date parser to see if it less then the date. Here's an expression to get you started:
(\d{4}-\d{2}-\d{2})\s+(.*)
Then the date will be in capture group one. If it is <= Groundhog's day, then you have the value in capture group two.
To show how complicated it is to do <= logic with regular expression, I whipped together a quick expression to match numbers > 0 and <= 27.
^([1-9]|1[0-9]|2[0-7])$
As you can see, we pretty much need to map out each scenario. You can imagine how much more of a headache this would be with a date..and you wouldn't just be able to say "2014-02-02", you'd need to redo the majority of the expression.

Get one letter abbreviation of week day of a date in java

As far I know, in Java I can get weekdays in normal (Friday) or short mode (Fri). But, there is any way to obtain only first letter?
I thought I can get first letter using "substring", but it won't be correct for all languages. For example, spanish weekdays are: Lunes, Martes, Miércoles, Jueves, Viernes, Sábado and Domingo, and first letter for "Miércoles" is X instead of M to difference it from "Martes".

In Android you can use SimpleDateFormat with "EEEEE". In the next example you can see it.
SimpleDateFormat formatLetterDay = new SimpleDateFormat("EEEEE",Locale.getDefault());
String letter = formatLetterDay.format(new Date());
EDIT: it's actually not entirely true. The result on Android could have more than a single letter (and also non-unique, if this matters), but this is what we have. Here's proof that you won't get these characteristics on Android, going over all locales. It's written in Kotlin, but should work for Java too, of course:
val charCountStats = SparseIntArray()
Locale.getAvailableLocales().forEach { locale ->
val sb = StringBuilder("$locale : ")
val formatLetterDay = SimpleDateFormat("EEEEE", locale)
for (day in 1..7) {
val cal = Calendar.getInstance()
cal.set(Calendar.DAY_OF_WEEK, day)
val letter: String = formatLetterDay.format(cal.time)
charCountStats.put(letter.length, charCountStats.get(letter.length, 0)+1)
sb.append(letter)
if (day != 7)
sb.append(',')
}
Log.d("AppLog", "$sb")
}
Log.d("AppLog", "stats:")
charCountStats.forEach { key, value ->
Log.d("AppLog", "formatted days with $key characters:$value")
}
And the result is that for most cases it's indeed a single letter, but for many it's more, and can even reaches 8 characters (though it might look as less letters, even one) :
formatted days with 1 characters:4889
formatted days with 2 characters:471
formatted days with 3 characters:99
formatted days with 4 characters:58
formatted days with 5 characters:3
formatted days with 8 characters:3
Example of a locale that it shows as 3 letters (and not just has 3 letters) is "wo" ("Wolof" language), as this is the result for each of its days of the week using the above formatting:
Dib,Alt,Tal,Àla,Alx,Àjj,Ase

As mentioned above there is no standard Java support for this. Using the formatting string "EEEEE" however is not guaranteed to work on all Android devices. The following code is guaranteed to work on any device:
public String firstLetterOfDayOfTheWeek(Date date) {
Locale locale = Locale.getDefault();
DateFormat weekdayNameFormat = new SimpleDateFormat("EEE", locale);
String weekday = weekdayNameFormat.format(date);
return weekday.charAt(0)+"";
}

There is no standard Java API support for doing that1.
Part of the reason is that many (maybe even most) languages don't have conventional unique one-letter weekday abbreviations. In English there isn't, for example (M T W T F S S).
A (hypothetical) formatting option that doesn't work2 in many / most locales would be an impediment to internationalization rather than a help.
It has been pointed out that:
SimpleDateFormat formatLetterDay =
new SimpleDateFormat("EEEEE", Locale.getDefault());
String letter = formatLetterDay.format(new Date());
gives one letter abbreviations for later versions of Android (18 and above), though the javadocs do not mention this. It appears that this "5 letter" format has been borrowed from DateTimeFormatter whose javadoc says:
The count of pattern letters determines the format.
Text: The text style is determined based on the number of pattern letters used. Less than 4 pattern letters will use the short form. Exactly 4 pattern letters will use the full form. Exactly 5 pattern letters will use the narrow form. ...
If you are targeting Android API 26 or later, you should consider using the java.time.* classes rather than the legacy classes.
But either way, this isn't guaranteed to give you unique day letters.
1 - By "that" I mean mapping to unique 1-letter abbreviations.
2 - I mean it doesn't work in the human sense. You could invent a convention, but typical people wouldn't understand what the abbreviations meant; e.g. they wouldn't know that "X" meant "Miércoles", or in English that (say) "R" meant "Thursday" (see https://stackoverflow.com/a/21049169/139985).

I realize the OP was asking for standards across languages, and this does not address it. But there is/was a standard for using single character Day of Week abbreviation.
Back in mainframe days, using a 1-character abbreviation for Day of Week was common, either to store day of week in 1 character field (to save precious space), or have a report heading for single-character column. The "standard" was to use MTWRFSU, where R was for Thursday, and U for Sunday.
I could not find any definitive references to this (which is why I quoted "standard", but here are a couple of examples:
http://eventguide.com/topics/one_digit_day_abbreviations.html
http://www.registrar.ucla.edu/soc/definitions.htm#Anchor-Days-3800

I think there's no direct java function to get the first letter and no standard way to do it.
You can refer to this link to obtain the first letter of the string day using substring() java method
Given a string in Java, just take the first X letters

DateFormatSymbols.getWeekdays with a width of NARROW will give you the first letter of each week day. It works for every language. However, it requires API 24.
if (Build.VERSION.SDK_INT >= Build.VERSION_CODES.N) {
String[] weekDays = DateFormatSymbols.getInstance(Locale.getDefault())
.getWeekdays(DateFormatSymbols.STANDALONE, DateFormatSymbols.NARROW);
}

Working with GPS data in Java

I'm using a GPS web service that is retrieving information in the following format (numbers changed up a bit for privacy reasons but format is unchanged):
X: 32 14 08.47S
Y: 140 17 12.82E
What I need to do is convert these to decimal co-ordinates (xx.xxxxxxxxx, xx.xxxxxxxxx). Are there any simple snippets of Java code that can do this task? If not, I'm happy to look at resources that explain how to achieve this in a different language.

If the degrees, minutes, and seconds are guaranteed to be separated a single space you could do something as simple as
String line = read_a_line_from_file();
String[] tokens = line.split(" ");
That will leave you with
tokens[0] = "X:"
tokens[1] = "32"
tokens[2] = "14"
tokens[3] = "08.47S"
You could then Double.parseDouble() the ones after tokens[0] to get the numeric degrees, minutes, and seconds which you would then combine to get the decimal degrees. Of course for tokens[3] you'd have to strip off the final N/S/E/W character before doing the parse.
Another more elegant possibility might be to take advantage of the fact that instances of MessageFormat and its subclasses can be use for parsing a string of a given format as well as formatting such a string.