Question

I’m attempting to parse a file that has a time stamp that i believe is 8 bytes long and looks like so in hex

00 00 00 00 DE A4 4F 4F

I do not receive the correct date/time when i parse this as an Int64. However if i skip the first 4 bytes and do something like so i get the correct datetime.

TimeSpan span = TimeSpan.FromTicks(BitConverter.ToInt32(bytes.Skip(index).Take(8).ToArray(),4) * TimeSpan.TicksPerSecond);
DateTime t = new DateTime(1970, 1, 1).Add(span);
StartTime = TimeZone.CurrentTimeZone.ToLocalTime(t);

However I’m not certain the next files i get to parse are going to have leading 00’s for the first 4 bytes. If i parse this as a ToInt64 i throw an outOfRange exception. What is the proper way to parse this?

Answer 1

Are you sure the values aren’t actually

DE A4 4F 4F 00 00 00 00

… for example, due to you miscounting the position in the file?

That would make more sense, and represent a little-endian 64-bit integer, which you’d convert like so:

static readonly DateTime Epoch = new DateTime(1970, 1, 1, 0, 0, 0,
                                              DateTimeKind.Utc);

...

long seconds = BitConverter.ToInt64(data, index);
DateTime date = Epoch + TimeSpan.FromSeconds(seconds);

(Note how much simpler this conversion code is, by the way – no messing around with LINQ, and no mulitplying the value by TicksPerSecond.)

EDIT: If it’s really two little-endian 32-bit integers, arranged in big-endian order to create a 64-bit integer, then you could use:

ulong x = BitConverter.ToUInt32(data, index);
ulong y = BitConverter.ToUInt32(data, index + 4);
ulong combined = (x << 32) | y;
long seconds = unchecked ((long) combined);