I'm writing a low level tool to scan and recover data from damaged Jet4 MDB files. I'm scanning through pages, parsing Rows, and decoding columns.
If I have the raw 8 byte value for Datetime fields, how can I convert this to a string representation of the date such as "MM-DD-YY HH:MM:SS"?
If I have the raw 4 byte value for a Number field with Single field size and 3 decimal places, how can I convert this value to a float/double?
Are there any documents that describe how all of the access fields are encoded and stored on disk?
Thank you.
If I have the raw 8 byte value for Datetime fields, how can I convert this to a string representation of the date such as "MM-DD-YY HH:MM:SS"?
Access stores Date/Time values as 64 bit (8 byte) Double
values in little-endian format. The integer portion represents the number of days before or after the Access "epoch" (1899-12-30 00:00:00
) and the absolute value of the fractional portion represents the time portion for that day (e.g., 0.5 = Noon). So, for example, in Python we would convert the bytes into a datetime
value like so:
from datetime import datetime, timedelta
import struct
# bytes as retrieved from .accdb or .mdb file
d_as_bytes = b'\x35\x07\x2F\x2C\x93\x63\xDD\xC0'
d_as_double = struct.unpack('<d', d_as_bytes)[0] # -30094.29957175926
d_integer_part = int(d_as_double) # -30094
d_fractional_part = abs(d_as_double - d_integer_part) # 0.29957175926
access_epoch = datetime(1899, 12, 30)
d = access_epoch + timedelta(days=d_integer_part) + timedelta(days=d_fractional_part)
print(d) # 1817-08-08 07:11:23
No. The JET format is proprietary of Microsoft and not documented.
As for the Date data type, this is simply a double.
Neither double nor single carry a format. As you mention, these a standard floating point values.
来源:https://stackoverflow.com/questions/45560782/how-are-datetime-and-number-formats-encoded-in-jet4