There are quite a few different conventions for binary datetime, depending on different
platforms and protocols. Some of these have severe drawbacks. For example, people using
Unix time (seconds since Jan 1, 1970) think that they are safe until near the year 2038.
But cases can and do arise where arithmetic manipulations causes serious problems. Consider
the computation of the average of two datetimes, for example: if one calculates them with
averageTime = (time1 + time2)/2
, there will be overflow even with dates
around the present. Moreover, even if these problems don't occur, there is the issue of
conversion back and forth between different systems.
Binary datetimes differ in a number of ways: the datatype, the unit,
and the epoch (origin). We'll refer to these as time scales. For example:
Table 1: Binary Time Scales
Source | Datatype | Unit | Epoch |
---|
JAVA_TIME | long | milliseconds | Jan 1, 1970 |
UNIX_TIME | int or long | seconds | Jan 1, 1970 |
ICU4C | double | milliseconds | Jan 1, 1970 |
WINDOWS_FILE_TIME | long | ticks (100 nanoseconds) | Jan 1, 1601 |
DOTNET_DATE_TIME | long | ticks (100 nanoseconds) | Jan 1, 0001 |
MAC_OLD_TIME | int | seconds | Jan 1, 1904 |
MAC_TIME | double | seconds | Jan 1, 2001 |
EXCEL_TIME | ? | days | Dec 31, 1899 |
DB2_TIME | ? | days | Dec 31, 1899 |
All of the epochs start at 00:00 am (the earliest possible time on the day in question),
and are assumed to be UTC.
The ranges for different datatypes are given in the following table (all values in years).
The range of years includes the entire range expressible with positive and negative
values of the datatype. The range of years for double is the range that would be allowed
without losing precision to the corresponding unit.
Units | long | double | int |
---|
1 sec | 5.84542×10¹¹ | 285,420,920.94 | 136.10 |
1 millisecond | 584,542,046.09 | 285,420.92 | 0.14 |
1 microsecond | 584,542.05 | 285.42 | 0.00 |
100 nanoseconds (tick) | 58,454.20 | 28.54 | 0.00 |
1 nanosecond | 584.5420461 | 0.2854 | 0.00 |
This class implements a universal time scale which can be used as a 'pivot',
and provide conversion functions to and from all other major time scales.
This datetimes to be converted to the pivot time, safely manipulated,
and converted back to any other datetime time scale.
So what to use for this pivot? Java time has plenty of range, but cannot represent
.NET framework
System.DateTime
vaules without severe loss of precision. ICU4C time addresses this by using a
double
that is otherwise equivalent to the Java time. However, there are disadvantages
with
doubles
. They provide for much more graceful degradation in arithmetic operations.
But they only have 53 bits of accuracy, which means that they will lose precision when
converting back and forth to ticks. What would really be nice would be a
long double
(80 bits -- 64 bit mantissa), but that is not supported on most systems.
The Unix extended time uses a structure with two components: time in seconds and a
fractional field (microseconds). However, this is clumsy, slow, and
prone to error (you always have to keep track of overflow and underflow in the
fractional field).
BigDecimal
would allow for arbitrary precision and arbitrary range,
but we would not want to use this as the normal type, because it is slow and does not
have a fixed size.
Because of these issues, we ended up concluding that the .NET framework's
System.DateTime
would be the
best pivot. However, we use the full range allowed by the datatype, allowing for
datetimes back to 29,000 BC and up to 29,000 AD. This time scale is very fine grained,
does not lose precision, and covers a range that will meet almost all requirements.
It will not handle the range that Java times would, but frankly, being able to handle dates
before 29,000 BC or after 29,000 AD is of very limited interest. However, for those cases,
we also allow conversion to an optional
BigDecimal
format that would have arbitrary
precision and range.
DB2_TIME
public static final int DB2_TIME
Used in DB2. Data is a ?unknown?
. Value
is days since December 31, 1899.
DOTNET_DATE_TIME
public static final int DOTNET_DATE_TIME
Used in the .NET framework's System.DateTime
structure.
Data is a long
. Value is ticks (1 tick == 100 nanoseconds) since January 1, 0001.
EPOCH_OFFSET_MINUS_1_VALUE
public static final int EPOCH_OFFSET_MINUS_1_VALUE
The constant used to select the epoch offset minus one value
for a time scale.
NOTE: This is an internal value. DO NOT USE IT. May not
actually be equal to the epoch offset value minus one.
EPOCH_OFFSET_PLUS_1_VALUE
public static final int EPOCH_OFFSET_PLUS_1_VALUE
The constant used to select the epoch plus one value
for a time scale.
NOTE: This is an internal value. DO NOT USE IT. May not
actually be equal to the epoch offset value plus one.
EPOCH_OFFSET_VALUE
public static final int EPOCH_OFFSET_VALUE
The constant used to select the epoch offset value
for a time scale.
EXCEL_TIME
public static final int EXCEL_TIME
Used in Excel. Data is a ?unknown?
. Value
is days since December 31, 1899.
FROM_MAX_VALUE
public static final int FROM_MAX_VALUE
The constant used to select the maximum from value
for a time scale.
FROM_MIN_VALUE
public static final int FROM_MIN_VALUE
The constant used to select the minimum from value
for a time scale.
ICU4C_TIME
public static final int ICU4C_TIME
Used in the ICU4C. Data is a double
. Value
is milliseconds since January 1, 1970.
JAVA_TIME
public static final int JAVA_TIME
Used in the JDK. Data is a long
. Value
is milliseconds since January 1, 1970.
MAC_OLD_TIME
public static final int MAC_OLD_TIME
Used in older Macintosh systems. Data is an int
. Value
is seconds since January 1, 1904.
MAC_TIME
public static final int MAC_TIME
Used in the JDK. Data is a double
. Value
is milliseconds since January 1, 2001.
MAX_ROUND_VALUE
public static final int MAX_ROUND_VALUE
The constant used to select the maximum safe rounding value
for a time scale.
NOTE: This is an internal value. DO NOT USE IT.
MAX_SCALE
public static final int MAX_SCALE
This is the first unused time scale value.
MAX_SCALE_VALUE
public static final int MAX_SCALE_VALUE
The number of time scale values.
NOTE: This is an internal value. DO NOT USE IT.
MIN_ROUND_VALUE
public static final int MIN_ROUND_VALUE
The constant used to select the minimum safe rounding value
for a time scale.
NOTE: This is an internal value. DO NOT USE IT.
TO_MAX_VALUE
public static final int TO_MAX_VALUE
The constant used to select the maximum to value
for a time scale.
TO_MIN_VALUE
public static final int TO_MIN_VALUE
The constant used to select the minimum to value
for a time scale.
UNITS_ROUND_VALUE
public static final int UNITS_ROUND_VALUE
The constant used to select the units round value
for a time scale.
NOTE: This is an internal value. DO NOT USE IT.
UNITS_VALUE
public static final int UNITS_VALUE
The constant used to select the units value
for a time scale.
UNIX_TIME
public static final int UNIX_TIME
Used in Unix systems. Data is an int
or a long
. Value
is seconds since January 1, 1970.
WINDOWS_FILE_TIME
public static final int WINDOWS_FILE_TIME
Used in Windows for file times. Data is a long
. Value
is ticks (1 tick == 100 nanoseconds) since January 1, 1601.
bigDecimalFrom
public static BigDecimal bigDecimalFrom(BigDecimal otherTime,
int timeScale)
Convert a BigDecimal
datetime from the given time scale to the universal time scale.
All calculations are done using BigDecimal
to guarantee that the value
does not go out of range.
otherTime
- The BigDecimal
datetimetimeScale
- The time scale to convert from
- The datetime converted to the universal time scale
bigDecimalFrom
public static BigDecimal bigDecimalFrom(double otherTime,
int timeScale)
Convert a double
datetime from the given time scale to the universal time scale.
All calculations are done using BigDecimal
to guarantee that the value
does not go out of range.
otherTime
- The double
datetimetimeScale
- The time scale to convert from
- The datetime converted to the universal time scale
bigDecimalFrom
public static BigDecimal bigDecimalFrom(long otherTime,
int timeScale)
Convert a long
datetime from the given time scale to the universal time scale.
All calculations are done using BigDecimal
to guarantee that the value
does not go out of range.
otherTime
- The long
datetimetimeScale
- The time scale to convert from
- The datetime converted to the universal time scale
from
public static long from(long otherTime,
int timeScale)
Convert a long
datetime from the given time scale to the universal time scale.
otherTime
- The long
datetimetimeScale
- The time scale to convert from
- The datetime converted to the universal time scale
getTimeScaleValue
public static long getTimeScaleValue(int scale,
int value)
Get a value associated with a particular time scale.
scale
- - the time scalevalue
- - a constant representing the value to get
toBigDecimal
public static BigDecimal toBigDecimal(BigDecimal universalTime,
int timeScale)
Convert a datetime from the universal time scale to a BigDecimal
in the given time scale.
universalTime
- The datetime in the universal time scaletimeScale
- The time scale to convert to
- The datetime converted to the given time scale
toBigDecimal
public static BigDecimal toBigDecimal(long universalTime,
int timeScale)
Convert a datetime from the universal time scale to a BigDecimal
in the given time scale.
universalTime
- The datetime in the universal time scaletimeScale
- The time scale to convert to
- The datetime converted to the given time scale
toBigDecimalTrunc
public static BigDecimal toBigDecimalTrunc(BigDecimal universalTime,
int timeScale)
Convert a time in the Universal Time Scale into another time
scale. The division used to do the conversion rounds down.
NOTE: This is an internal routine used by the tool that
generates the to and from limits. Use it at your own risk.
universalTime
- the time in the Universal Time scaletimeScale
- the time scale to convert to
- the time in the given time scale
toLong
public static long toLong(long universalTime,
int timeScale)
Convert a datetime from the universal time scale stored as a
BigDecimal
to a
long
in the given time scale.
Since this calculation requires a divide, we must round. The straight forward
way to round by adding half of the divisor will push the sum out of range for values
within have the divisor of the limits of the precision of a
long
. To get around this, we do
the rounding like this:
(universalTime - units + units/2) / units + 1
(i.e. we subtract units first to guarantee that we'll still be in range when we
add
units/2
. We then need to add one to the quotent to make up for the extra subtraction.
This simplifies to:
(universalTime - units/2) / units - 1
For negative values to round away from zero, we need to flip the signs:
(universalTime + units/2) / units + 1
Since we also need to subtract the epochOffset, we fold the
+/- 1
into the offset value. (i.e.
epochOffsetP1
,
epochOffsetM1
.)
universalTime
- The datetime in the universal time scaletimeScale
- The time scale to convert to
- The datetime converted to the given time scale