I need to read binary files which consist of 19 float32 numbers followed by a unknown number of uint32 numbers. How can I read such a file in python?
In Matlab the equivalent looks like this:
fid = fopen('myFile.bin','r');
params = fread(fid,19,'float');
data = fread(fid,'uint32');
fclose(fid);
Use numpy.fromfile() method and pass a file handle to it with the corresponding number of items to read.
import numpy as np
with open('myFile.bin', 'rb') as f:
params = np.fromfile(f, dtype=np.float32, count=19)
data = np.fromfile(f, dtype=np.int32, count=-1) # I *assumed* here your ints are 32-bit
Postpend .tolist() to closing paranthesis of fromfile() (like this: np.fromfile(...).tolist()) if you want to get standard Python lists instead of numpy arrays.
counts is set to -1 by default and that is why I did not explicitly put it in the second call. From docs: count : int Number of items to read. -1 means all items (i.e., the complete file). and numpy.fromfile(file, dtype=float, count=-1, sep='') So, technically, last -1 in the second fromfile() is not required to read the rest of integer numbers.
For reading binary file I recomend using struct package
The solution can be written as follows:
import struct
f = open("myFile.bin", "rb")
floats_bytes = f.read(19 * 4)
# here I assume that we read exactly 19 floats without errors
# 19 floats in array
floats_array = struct.unpack("<19f", floats_bytes)
# convert to list beacause struct.unpack returns tuple
floats_array = list(floats_array)
# array of ints
ints_array = []
while True:
int_bytes = r.read(4)
# check if not eof
if not int_bytes:
break
int_value = struct.unpack("<I", int_bytes)[0]
ints_array.append(int_value)
f.close()
Note that I assumed your numbers are stored in little endian byte order, so I used "<" in format strings.
