Hello, all. Thank you for your time. I am a new user, on the 2012 Windows version.
The short, broad question: is there any way to work with a binary file bit-wise, instead of byte-wise? Or, alternately, to use a custom data type with "read from binary file," which is not a full byte long?
The explanation:
I have data from, in essence, a counter. It dumps binary files in which each event has information encoded in a word of N bytes. However, subsets of information are not necessarily a full byte, nor are they "aligned" to bytes. (For the record, I can't change the data type)
A concrete example, slightly simplified but effectively identical: A word length is 2 bytes. In this word, the first 5 bits are an ID number, and the remaining 11 bits are a timestamp.
01101 11111000000
ID Timestamp
I am trying to quickly parse this file, creating an array with the information.
I can do it in the following manner:
-Read data in an intermediate format (i.e. U16)
-Convert that format to binary
-Chop the binary string to the desired lengths
-Convert back to integer
However, this is prohibitively slow (for essentially real-time processing), because of the necessity to loop through a large array. There are several potential solutions, but I haven't been able to implement any. Suggestions on their implementation? Or other solutions?
-Create custom data type for "read from binary file." Hasn't worked thus far because even a sub-byte read will result in a byte change in file position, or else the necessary conversions take as long as the original solution.
-Convert entire file into a 2D binary array, split array, convert array to numbers. Can't get booleans to read from bits instead of bytes.
Your help is much appreciated!
