seek() function?

Regarding seek() there's not too much to worry about.

First of all, it is useful when operating over an open file.

It's important to note that its syntax is as follows:

fp.seek(offset, from_what)

where fp is the file pointer you're working with; offset means how many positions you will move; from_what defines your point of reference:

  • 0: means your reference point is the beginning of the file
  • 1: means your reference point is the current file position
  • 2: means your reference point is the end of the file

if omitted, from_what defaults to 0.

Never forget that when managing files, there'll always be a position inside that file where you are currently working on. When just open, that position is the beginning of the file, but as you work with it, you may advance.
seek will be useful to you when you need to walk along that open file, just as a path you are traveling into.


The seek function expect's an offset in bytes.

Ascii File Example:

So if you have a text file with the following content:

simple.txt

abc

You can jump 1 byte to skip over the first character as following:

fp = open('simple.txt', 'r')
fp.seek(1)
print fp.readline()
>>> bc

Binary file example gathering width :

fp = open('afile.png', 'rb')
fp.seek(16)
print 'width: {0}'.format(struct.unpack('>i', fp.read(4))[0])
print 'height: ', struct.unpack('>i', fp.read(4))[0]

Note: Once you call read you are changing the position of the read-head, which act's like seek.


When you open a file, the system points to the beginning of the file. Any read or write you do will happen from the beginning. A seek() operation moves that pointer to some other part of the file so you can read or write at that place.

So, if you want to read the whole file but skip the first 20 bytes, open the file, seek(20) to move to where you want to start reading, then continue with reading the file.

Or say you want to read every 10th byte, you could write a loop that does seek(9, 1) (moves 9 bytes forward relative to the current positions), read(1) (reads one byte), repeat.

Tags:

Python