I have large binary data files that have a predefined format, originally written by a Fortran program as little endians. I would like to read these files in the fastest, most efficient manner, so using the array package seemed right up my alley as suggested in Improve speed of reading and converting from binary file?.
The problem is the pre-defined format is non-homogeneous. It looks something like this:
['<2i','<5d','<2i','<d','<i','<3d','<2i','<3d','<i','<d','<i','<3d']
with each integer i taking up 4 bytes, and each double d taking 8 bytes.
Is there a way I can still use the super efficient array package (or another suggestion) but with the right format?
Use struct. In particular, struct.unpack.
result = struct.unpack("<2i5d...", buffer)
Here buffer holds the given binary data.
It's not clear from your question whether you're concerned about the actual file reading speed (and building data structure in memory), or about later data processing speed.
If you are reading only once, and doing heavy processing later, you can read the file record by record (if your binary data is a recordset of repeated records with identical format), parse it with struct.unpack and append it to a [double] array:
from functools import partial
data = array.array('d')
record_size_in_bytes = 9*4 + 16*8 # 9 ints + 16 doubles
with open('input', 'rb') as fin:
for record in iter(partial(fin.read, record_size_in_bytes), b''):
values = struct.unpack("<2i5d...", record)
data.extend(values)
Under assumption you are allowed to cast all your ints to doubles and willing to accept increase in allocated memory size (22% increase for your record from the question).
If you are reading the data from file many times, it could be worthwhile to convert everything to one large array of doubles (like above) and write it back to another file from which you can later read with array.fromfile():
data = array.array('d')
with open('preprocessed', 'rb') as fin:
n = os.fstat(fin.fileno()).st_size // 8
data.fromfile(fin, n)
Update. Thanks to a nice benchmark by #martineau, now we know for a fact that preprocessing the data and turning it into an homogeneous array of doubles ensures that loading such data from file (with array.fromfile()) is ~20x to ~40x faster than reading it record-per-record, unpacking and appending to array (as shown in the first code listing above).
A faster (and a more standard) variation of record-by-record reading in #martineau's answer which appends to list and doesn't upcast to double is only ~6x to ~10x slower than array.fromfile() method and seems like a better reference benchmark.
Major Update: Modified to use proper code for reading in a preprocessed array file (function using_preprocessed_file() below), which dramatically changed the results.
To determine what method is faster in Python (using only built-ins and the standard libraries), I created a script to benchmark (via timeit) the different techniques that could be used to do this. It's a bit on the longish side, so to avoid distraction, I'm only posting the code tested and related results. (If there's sufficient interest in the methodology, I'll post the whole script.)
Here are the snippets of code that were compared:
#TESTCASE('Read and constuct piecemeal with struct')
def read_file_piecemeal():
structures = []
with open(test_filenames[0], 'rb') as inp:
size = fmt1.size
while True:
buffer = inp.read(size)
if len(buffer) != size: # EOF?
break
structures.append(fmt1.unpack(buffer))
return structures
#TESTCASE('Read all-at-once, then slice and struct')
def read_entire_file():
offset, unpack, size = 0, fmt1.unpack, fmt1.size
structures = []
with open(test_filenames[0], 'rb') as inp:
buffer = inp.read() # read entire file
while True:
chunk = buffer[offset: offset+size]
if len(chunk) != size: # EOF?
break
structures.append(unpack(chunk))
offset += size
return structures
#TESTCASE('Convert to array (#randomir part 1)')
def convert_to_array():
data = array.array('d')
record_size_in_bytes = 9*4 + 16*8 # 9 ints + 16 doubles (standard sizes)
with open(test_filenames[0], 'rb') as fin:
for record in iter(partial(fin.read, record_size_in_bytes), b''):
values = struct.unpack("<2i5d2idi3d2i3didi3d", record)
data.extend(values)
return data
#TESTCASE('Read array file (#randomir part 2)', setup='create_preprocessed_file')
def using_preprocessed_file():
data = array.array('d')
with open(test_filenames[1], 'rb') as fin:
n = os.fstat(fin.fileno()).st_size // 8
data.fromfile(fin, n)
return data
def create_preprocessed_file():
""" Save array created by convert_to_array() into a separate test file. """
test_filename = test_filenames[1]
if not os.path.isfile(test_filename): # doesn't already exist?
data = convert_to_array()
with open(test_filename, 'wb') as file:
data.tofile(file)
And here were the results running them on my system:
Fastest to slowest execution speeds using Python 3.6.1
(10 executions, best of 3 repetitions)
Size of structure: 164
Number of structures in test file: 40,000
file size: 6,560,000 bytes
Read array file (#randomir part 2): 0.06430 secs, relative 1.00x ( 0.00% slower)
Read all-at-once, then slice and struct: 0.39634 secs, relative 6.16x ( 516.36% slower)
Read and constuct piecemeal with struct: 0.43283 secs, relative 6.73x ( 573.09% slower)
Convert to array (#randomir part 1): 1.38310 secs, relative 21.51x (2050.87% slower)
Interestingly, most of the snippets are actually faster in Python 2...
Fastest to slowest execution speeds using Python 2.7.13
(10 executions, best of 3 repetitions)
Size of structure: 164
Number of structures in test file: 40,000
file size: 6,560,000 bytes
Read array file (#randomir part 2): 0.03586 secs, relative 1.00x ( 0.00% slower)
Read all-at-once, then slice and struct: 0.27871 secs, relative 7.77x ( 677.17% slower)
Read and constuct piecemeal with struct: 0.40804 secs, relative 11.38x (1037.81% slower)
Convert to array (#randomir part 1): 1.45830 secs, relative 40.66x (3966.41% slower)
Take a look at the documentation for numpy's fromfile function: https://docs.scipy.org/doc/numpy-1.13.0/reference/generated/numpy.fromfile.html and https://docs.scipy.org/doc/numpy/reference/arrays.dtypes.html#arrays-dtypes-constructing
Simplest example:
import numpy as np
data = np.fromfile('binary_file', dtype=np.dtype('<i8, ...'))
Read more about "Structured Arrays" in numpy and how to specify their data type(s) here: https://docs.scipy.org/doc/numpy/user/basics.rec.html#
There's a lot of good and helpful answers here, but I think the best solution needs more explaining. I implemented a method that reads the entire data file in one pass using the built-in read() and constructs a numpy ndarray all at the same time. This is more efficient than reading the data and constructing the array separately, but it's also a bit more finicky.
line_cols = 20 #For example
line_rows = 40000 #For example
data_fmt = 15*'f8,'+5*'f4,' #For example (15 8-byte doubles + 5 4-byte floats)
data_bsize = 15*8 + 4*5 #For example
with open(filename,'rb') as f:
data = np.ndarray(shape=(1,line_rows),
dtype=np.dtype(data_fmt),
buffer=f.read(line_rows*data_bsize))[0].astype(line_cols*'f8,').view(dtype='f8').reshape(line_rows,line_cols)[:,:-1]
Here, we open the file as a binary file using the 'rb' option in open. Then, we construct our ndarray with the proper shape and dtype to fit our read buffer. We then reduce the ndarray into a 1D array by taking its zeroth index, where all our data is hiding. Then, we reshape the array using np.astype, np.view and np.reshape methods. This is because np.reshape doesn't like having data with mixed dtypes, and I'm okay with having my integers expressed as doubles.
This method is ~100x faster than looping line-for-line through the data, and could potentially be compressed down into a single line of code.
In the future, I may try to read the data in even faster using a Fortran script that essentially converts the binary file into a text file. I don't know if this will be faster, but it may be worth a try.
Related
Maybe there's a way around this that I'm missing. Long story short, I have a need for shared memory access, read only, of a large text file. Working with strings of course is necessary. So I'm trying to do this:
import numpy
from multiprocessing import Pool, RawArray
if __name__ == '__main__':
with open('test.txt', 'r') as fin:
raw = fin.readlines()
X_shape = (len(raw), 70) # 70 characters per line should be sufficient for my needs
X = RawArray('c', X_shape[0] * X_shape[1])
X_np = np.frombuffer(X).reshape(X_shape)
numpy.copyto(X_np, raw)
This doesn't work, it fails on the second last line with this output:
ValueError: cannot reshape array of size 102242175 into shape (11684820,70)
For reference the file sample is 11684820 lines long. And 11684820 * 70 is definitely not going to be the number of characters the array claims it is sized for.
Clearly I must be doing something wrong, but this is the only method I see as feasible to multiprocess some CPU bound computations using text file inputs of text files that are several hundred megabytes on the low end and around 6 gigabytes on the high end.
Is there a work around, or perhaps a more correct way of doing this so I can have a large array of strings in shared memory that I can work on with python code? Thanks.
numpy.frombuffer needs an explicit dtype, or it will default to dtype=float. Also, a 11684820x70 array of uint8s or 1-character bytestrings isn't the same as a length-11684820 array of 70-character bytestrings, so keep that in mind.
For a 11684820x70 array, the shape you asked for, but probably not what you need:
X_np = np.frombuffer(X, dtype=np.uint8).reshape(X_shape)
For a length-11684820 array of dtype S70 (null-terminated bytestrings of max length 70, described as "not recommended" in the NumPy docs):
X_np = np.frombuffer(X, dtype='S70')
For a length-11684820 array of dtype U70 (null-terminated Unicode strings of max length 70), you'll need a bigger buffer (4 bytes per character), and then
X_np = np.frombuffer(X, dtype='U70')
I have some very large txt fils(about 1.5 GB ) which I want to load into Python as an array. The Problem is in this data a comma is used as a decimal separator. for smaller fils I came up with this solution:
import numpy as np
data= np.loadtxt(file, dtype=np.str, delimiter='\t', skiprows=1)
data = np.char.replace(data, ',', '.')
data = np.char.replace(data, '\'', '')
data = np.char.replace(data, 'b', '').astype(np.float64)
But for the large fils Python runs into an Memory Error. Is there any other more memory efficient way to load this data?
The problem with np.loadtxt(file, dtype=np.str, delimiter='\t', skiprows=1) is that it uses python objects (strings) instead of float64, which is very memory inefficient. You can use pandas read_table
http://pandas.pydata.org/pandas-docs/stable/generated/pandas.read_table.html#pandas.read_table
to read your file and set decimal=',' to change the default behaviour. This will allow for seamless reading and converting your strings into floats. After loading pandas dataframe use df.values to get a numpy array.
If it's still too large for your memory use chunks
http://pandas.pydata.org/pandas-docs/stable/io.html#io-chunking
If still no luck try np.float32 format which further halves memory footprint.
You should try parsing it yourself, iterating for each line (so implicitly using a generator that does not read all the file into memory).
Also, for data of that size, I would use python standard array library, that uses similar memory as a c array. That is, one value next to the other in memory (numpy array is also very efficient in memory usage though).
import array
def convert(s):
# The function that converts the string to float
s = s.strip().replace(',', '.')
return float(s)
data = array.array('d') #an array of type double (float of 64 bits)
with open(filename, 'r') as f:
for l in f:
strnumbers = l.split('\t')
data.extend( (convert(s) for s in strnumbers if s!='') )
#A generator expression here.
I'm sure similar code (with similar memory footprint) can be written replacing the array.array by numpy.array, specially if you need to have a two dimensional array.
I have a piece of software that reads a file and transforms each first value it reads per line using a function (derived from numpy.polyfit and numpy.poly1d functions).
This function has to then write the transformed file away and I wrongly (it seems) assumed that the disk I/O part was the performance bottleneck.
The reason why I claim that it is the transformation that is slowing things down is because I tested the code (listed below) after i changed transformedValue = f(float(values[0])) into transformedValue = 1000.00 and that took the time required down from 1 min to 10 seconds.
I was wondering if anyone knows of a more efficient way to perform repeated transformations like this?
Code snippet:
def transformFile(self, f):
""" f contains the function returned by numpy.poly1d,
inputFile is a tab seperated file containing two floats
per line.
"""
with open (self.inputFile,'r') as fr:
for line in fr:
line = line.rstrip('\n')
values = line.split()
transformedValue = f(float(values[0])) # <-------- Bottleneck
outputBatch.append(str(transformedValue)+" "+values[1]+"\n")
joinedOutput = ''.join(outputBatch)
with open(output,'w') as fw:
fw.write(joinedOutput)
The function f is generated by another function, the function fits a 2d degree polynomial through a set of expected floats and a set of measured floats. A snippet from that function is:
# Perform 2d degree polynomial fit
z = numpy.polyfit(measuredValues,expectedValues,2)
f = numpy.poly1d(z)
-- ANSWER --
I have revised the code to vectorize the values prior to transforming them, which significantly speed-up the performance, the code is now as follows:
def transformFile(self, f):
""" f contains the function returned by numpy.poly1d,
inputFile is a tab seperated file containing two floats
per line.
"""
with open (self.inputFile,'r') as fr:
outputBatch = []
x_values = []
y_values = []
for line in fr:
line = line.rstrip('\n')
values = line.split()
x_values.append(float(values[0]))
y_values.append(int(values[1]))
# Transform python list into numpy array
xArray = numpy.array(x_values)
newArray = f(xArray)
# Prepare the outputs as a list
for index, i in enumerate(newArray):
outputBatch.append(str(i)+" "+str(y_values[index])+"\n")
# Join the output list elements
joinedOutput = ''.join(outputBatch)
with open(output,'w') as fw:
fw.write(joinedOutput)
It's difficult to suggest improvements without knowing exactly what your function f is doing. Are you able to share it?
However, in general many NumPy operations often work best (read: "fastest") on NumPy array objects rather than when they are repeated multiple times on individual values.
You might like to consider reading the numbers values[0] into a Python list, passing this to a NumPy array and using vectorisable NumPy operations to obtain an array of output values.
I compare two simple methods for writing a numpy array into a raw binary file :
# method 1
import numpy
A = numpy.random.randint(1000, size=512*1024*1024) # 2 GB
with open('blah.bin', 'wb') as f:
f.write(A)
and
# method 2
import numpy
A = numpy.random.randint(1000, size=512*1024*1024) # 2 GB
raw_input()
B = A.tostring() # check memory usage of the current process here : 4 GB are used !!
raw_input()
with open('blah.bin', 'wb') as f:
f.write(B)
With the second method, the memory usage is doubled (4 GB here) !
Why is .tostring() often used for writing numpy arrays to file ?
(in http://docs.scipy.org/doc/numpy/reference/generated/numpy.ndarray.tofile.html, it is explained that numpy.ndarray.tofile() may be equivalent to file.write(a.tostring()))
Is method 1 as correct as method 2 for writing such an array to disk ?
The documentation does not say that .tofile() is equivalent to file.write(a.tostring()), it only mentions the latter to explain how the argument sep will behave if it's value is "".
In the second method you are creating a copy of the array A, storing in B, and after that you write in the file, while in the first method this intermediate copy is avoided.
You should also have a look in:
np.savetxt()
When I load an array using numpy.loadtxt, it seems to take too much memory. E.g.
a = numpy.zeros(int(1e6))
causes an increase of about 8MB in memory (using htop, or just 8bytes*1million \approx 8MB). On the other hand, if I save and then load this array
numpy.savetxt('a.csv', a)
b = numpy.loadtxt('a.csv')
my memory usage increases by about 100MB! Again I observed this with htop. This was observed while in the iPython shell, and also while stepping through code using Pdb++.
Any idea what's going on here?
After reading jozzas's answer, I realized that if I know ahead of time the array size, there is a much more memory efficient way to do things if say 'a' was an mxn array:
b = numpy.zeros((m,n))
with open('a.csv', 'r') as f:
reader = csv.reader(f)
for i, row in enumerate(reader):
b[i,:] = numpy.array(row)
Saving this array of floats to a text file creates a 24M text file. When you re-load this, numpy goes through the file line-by-line, parsing the text and recreating the objects.
I would expect memory usage to spike during this time, as numpy doesn't know how big the resultant array needs to be until it gets to the end of the file, so I'd expect there to be at least 24M + 8M + other temporary memory used.
Here's the relevant bit of the numpy code, from /lib/npyio.py:
# Parse each line, including the first
for i, line in enumerate(itertools.chain([first_line], fh)):
vals = split_line(line)
if len(vals) == 0:
continue
if usecols:
vals = [vals[i] for i in usecols]
# Convert each value according to its column and store
items = [conv(val) for (conv, val) in zip(converters, vals)]
# Then pack it according to the dtype's nesting
items = pack_items(items, packing)
X.append(items)
#...A bit further on
X = np.array(X, dtype)
This additional memory usage shouldn't be a concern, as this is just the way python works - while your python process appears to be using 100M of memory, internally it maintains knowledge of which items are no longer used, and will re-use that memory. For example, if you were to re-run this save-load procedure in the one program (save, load, save, load), your memory usage will not increase to 200M.
Here is what I ended up doing to solve this problem. It works even if you don't know the shape ahead of time. This performs the conversion to float first, and then combines the arrays (as opposed to #JohnLyon's answer, which combines the arrays of string then converts to float). This used an order of magnitude less memory for me, although perhaps was a bit slower. However, I literally did not have the requisite memory to use np.loadtxt, so if you don't have sufficient memory, then this will be better:
def numpy_loadtxt_memory_friendly(the_file, max_bytes = 1000000, **loadtxt_kwargs):
numpy_arrs = []
with open(the_file, 'rb') as f:
i = 0
while True:
print(i)
some_lines = f.readlines(max_bytes)
if len(some_lines) == 0:
break
vec = np.loadtxt(some_lines, **loadtxt_kwargs)
if len(vec.shape) < 2:
vec = vec.reshape(1,-1)
numpy_arrs.append(vec)
i+=len(some_lines)
return np.concatenate(numpy_arrs, axis=0)