parsing_data.rst
author amit
Wed, 22 Sep 2010 14:56:22 +0530
changeset 179 1d04b6c5ff44
parent 140 bc023595e167
child 197 97d859b70f51
permissions -rw-r--r--
First Review for parsing_data.rst

Hello friends and welcome to the tutorial on Parsing Data

{{{ Show the slide containing title }}}

{{{ Show the slide containing the outline slide }}}

In this tutorial, we shall learn

 * What is parsing data
 * the string operations required for parsing data
 * datatype conversion

 Lets us have a look at the problem

{{{ Show the slide containing problem statement. }}}

There is an input file containing huge no.of records. Each record corresponds
to a student.

{{{ show the slide explaining record structure }}}
As you can see, each record consists of fields seperated by a ";". The first
record is region code, then roll number, then name, marks of second language,
first language, maths, science and social, total marks, pass/fail indicatd by P
or F and finally W if with held and empty otherwise.

Our job is to calculate the mean of all the maths marks in the region "B".

#[Nishanth]: Please note that I am not telling anything about AA since they do
             not know about any if/else yet.


So what exactly is parsing data?


Parsing data is all about reading the data and converting it into a form which
can be used for computations. In our case, that is numbers.

We can clearly see that the problem involves reading files and tokenizing.

.. #[[Amit:Definition of Tokenizing here.]]
Let us learn about tokenizing strings. Let us define a string first. Type
::

    line = "parse this           string"

We are now going to split this string on whitespace.
::

    line.split()

As you can see, we get a list of strings. Which means, when split is called
without any arguments, it splits on whitespace. In simple words, all the spaces
are treated as one big space.

split also can split on a string of our choice. This is acheived by passing
that as an argument. But first lets define a sample record from the file.
::

    record = "A;015163;JOSEPH RAJ S;083;042;47;AA;72;244;;;"
    record.split(';')

We can see that the string is split on ';' and we get each field seperately.
We can also observe that an empty string appears in the list since there are
two semi colons without anything in between.

Hence split splits on whitespace if called without an argument and splits on
the given argument if it is called with an argument.

{{{ Pause here and try out the following exercises }}}

%% 1 %% split the variable line using a space as argument. Is it same as
        splitting without an argument ?

{{{ continue from paused state }}}

We see that when we split on space, multiple whitespaces are not clubbed as one
and there is an empty string everytime there are two consecutive spaces.

Now that we know how to split a string, we can split the record and retreive each
field seperately. But there is one problem. The region code "B" and a "B"
surrounded by whitespace are treated as two different regions. We must find a
way to remove all the whitespace around a string so that "B" and a "B" with
white spaces are dealt as same.

This is possible by using the =strip= method of strings. Let us define a
string by typing
::

    unstripped = "     B    "
    unstripped.strip()

We can see that strip removes all the whitespace around the sentence

{{{ Pause here and try out the following exercises }}}

%% 2 %% What happens to the white space inside the sentence when it is stripped

{{{ continue from paused state }}}

Type
::

    a_str = "         white      space            "
    a_str.strip()

We see that the whitespace inside the sentence is only removed and anything
inside remains unaffected.

By now we know enough to seperate fields from the record and to strip out any
white space. The only road block we now have is conversion of string to float.

The splitting and stripping operations are done on a string and their result is
also a string, hence the marks that we have are still strings and mathematical
operations on them are not possible. We must convert them into integers or floats

We shall look at converting strings into floats. We define an float string
first. Type
::

    mark_str = "1.25"
    mark = float(mark_str)
    type(mark_str)
    type(mark)

We can see that string is converted to float. We can perform mathematical
operations on it now.

{{{ Pause here and try out the following exercises }}}

%% 3 %% What happens if you do int("1.25")

{{{ continue from paused state }}}

.. #[[Amit:I think there should be some interaction first here about the
problem before we conclude to talking about the result.]]
It raises an error since converting a float string into integer directly is
not possible. It involves an intermediate step of converting to float.
::

    dcml_str = "1.25"
    flt = float(dcml_str)
    flt
    number = int(flt)
    number

Using =int= it is possible to convert float into integers.

Now that we have all the machinery required to parse the file, let us solve the
problem. We first read the file line by line and parse each record. We see if
the region code is B and store the marks accordingly.
::

    math_marks_B = [] # an empty list to store the marks
    for line in open("/home/fossee/sslc1.txt"):
        fields = line.split(";")

        region_code = fields[0]
        region_code_stripped = region_code.strip()

        math_mark_str = fields[5]
        math_mark = float(math_mark_str)

        if region_code == "AA":
            math_marks_B.append(math_mark)
.. #[[Amit:This intutively does not seem to be what you wanted]]

Now we have all the maths marks of region "B" in the list math_marks_B.
To get the mean, we just have to sum the marks and divide by the length.
::

        math_marks_mean = sum(math_marks_B) / len(math_marks_B)
        math_marks_mean

{{{ Show summary slide }}}

This brings us to the end of the tutorial.
we have learnt

 * how to tokenize a string using various delimiters
 * how to get rid of extra white space around
 * how to convert from one type to another
.. #[[Amit:one datatype to another may be better.]]
 * how to parse input data and perform computations on it

{{{ Show the "sponsored by FOSSEE" slide }}}

#[Nishanth]: Will add this line after all of us fix on one.
This tutorial was created as a part of FOSSEE project, NME ICT, MHRD India

Hope you have enjoyed and found it useful.
Thankyou
 
.. Author              : Nishanth
   Internal Reviewer 1 : Amit Sethi 
   Internal Reviewer 2 : 
   External Reviewer   :