So, in the previous session, we have learned how to list the packages. So, we have used a command called dpkg query, which allowed us to combine the various fields to output them onto the screen, and we have seen a list of packages that is coming onto the screen. And we have used the grep to pick up, and we can actually now modify that to a little bit further extent.
So, I am just going to run the same query, so minus W to provide information about the packages and minus f to provide the pattern. So, I have the section I would like to have and then one space, and then the name of the package has to be shown, and after that I would like to have the newline character, and then now you see that this list is quite long. And here is where we would like to use the grep features.
So, let us say I use the extended regular expression engine, and I would like to look at all those package names which have

exactly four characters. So, W get is a package 
name with four characters, n map is a package name   with four characters. So, I want to pick up those 
kinds of package names. So, how do I provide the   pattern? So, I would actually have space because 
the package name is coming after the space.
  So, I do not want to match those which are   in the section, only in the package I would   like to match. And then I do not know what 
character has to be matched, so I put a dot.   So, any character can be there, but that 
should occur exactly four times. And after   the package name has been given, I should go to   the end of the line. So, the dollar will tell me   that it is going to be end of the line.
So, now look at that and you see that we   are getting package names which are exactly four 
characters long. And what happens if I do not have

the dollar there. You see that it would match four characters, but the remaining characters are also present which are not being looked at. So, by putting the dollar here we are insisting that the package name should have been finished in just four characters. And what is the blank here for, so I would actually also show you what happens if you do not have the blank here. You would see that the matching is occurring also in the name of the section so that I do not want. So, I put a blank here so that it is matching from the second field onwards and then it is matching only in the name of the package and the dollar is actually matching only the end of the line to be occurring after four characters of the name of the package. Now, let us say I would like to have the name of the package to be starting with g and the three characters after that, so g something, something, something. Now, you see

that this is giving us the package names which 
starts with the g and are four later words, gmp, or grep, or gzip, or gftp, etc. And we could 
also look at, after g, I can have either 1, or up to let us say 5 characters. 
And you see that slightly a longer list of g, starting packages are there, g plus plus onwards 
all the way up to guake terminal, which we have looked at in one of the earlier classes. So, 
you could see that the egrep can be used in a very powerful manner to process the matching of 
the string precisely, telling what character you would like to match and how many times, etc. 
Now, let us say you would like to have all those packages which are coming under 
the category called mathematics. So, that you want to go through those packages 
and be thorough of it. So, let us look at that. Math should be matched within the name of the 
package. Now, you see that it is matching also

names of packages here, but not only sections. 
Now, I want to match it only for the section. So which means that I want the math character 
set to should be only in the beginning of the line. So already we know the anchor for 
that. So I put the hat in the front.
And now you see that only those packages 
which come under the section called math are being displayed, but not the names 
of packages which have math in middle but actually belonging to some other section. So, 
you see the utility of the egrep along with the features such as the hat as well as the number 
of times the particular character should come so that you can actually precisely pick up the 
lines of output that you are interested in.
Now, there are some packages which are coming 
from the kde graphical user interface of Ubuntu and the default GUI for Ubuntu is 
gnome. So, you can look at those also here. So,

k and after that I have something that starts with k and then a dollar. So, packages that starts with k. And there are many of them. So, I think that k is not sufficient, kd. So, now this is much better, kde is common desktop environment and there are some packages that are coming. So, kdenlive is like an editor used to trim the videos etc., which is what I do for our course. And there are also other tools. So, these are coming from a common desktop environment, which requires some other libraries also to be present compared to the gnome user interface. So, you think about what way you want to pick up and analyze the pattern that would capture them accurately. And if you have used the thermal colors enabled then you can also find out what has been matched and that would also help you understand the pattern matching better. Now, there are some features that are available

in the grep which help you with respect to the type of character sets that can be matched, so alphabetic, alphanumeric, digits etc. So, we will look at them using a file in which I have put intentionally some punctuation marks or control characters. How do we insert non-typable characters in a text file is something that we will learn when we come to the text editors shortly, but for now I will just display the text file to you and then we will see how we can pick specific lines from it depending upon the type of character set that we would like to match. So, here is a file that I have in place. So, character types, I have got lines in which we have got various strings. So, I have got a beginning of the line with something and ending also with something intentionally so that those strings will be matched appropriately. So, I have got a lower letter, lowercase letter, I have got uppercase letter, I have got a digit.

and then I also have a non-printable character, like Ctrl + C. And here also there is a tab character that has been used and so on. And there is a blank space here in front of it. So, and then there is an empty line also in the middle. So, let us see how we can pick these lines by using the character sets which come with a double square bracket and colons separating the name of the character set that we are using. So, in all of them I will use the same pattern where we will use a cat chartypes pipe grep. Now, let us use this pattern. Here we would like to let us say pick alphabetic characters. So, any line that would have an alphabetical character would be matched, and you see that all the lines will be matched, except the blank line, because the bank line has nothing in it, so that has been skipped.

Next, the alphanum, that is lines which would have alphanumeric characters will be matched. And you see that it is the same output as the previous one. And it has also skipped the empty line because it has no alphanumeric character that can be matched on that line. Now, let us say I would like to have the alphabetical character in the beginning of the string. Now, you see that only some of the lines are showing up. So, if you see hello and then l and capital L these are getting matched because they have an alphabetical character at the beginning, but the remaining lines do not have. It has 5 and it has a 4 and this fellow has a control c and there is a empty character here null and this is a comma here and there is a blank. So, none of them are actually alphabetical characters, so they are all skipped. Now, you could also ask for the alphabetical character to be at the end of the

line. And you will see that only some of the lines are being shown, the one with a g, capital H and a small h. The remaining lines are not shown because here it is a 2 which is a digit, and here it is a tab character and here it is a dot and here there is a null and here there is an equal to sign, and here again, there is a control character. So all of those have been skipped. And we have got only the first three lines which have an alphabetical character in the end. So, you now see that alphabetical character can be matched either at the beginning or anywhere in the line or at the end and you need to be watchful which ones are suitable for your purpose. Now, we can also look at the alphanumeric character at the beginning of the line. And you see that it captures a little bit more number of lines. So, the first character can

be either an alphabetical or a numeric one, but 
not a space or a null or a punctuation mark or a control character. So, those lines have been 
skipped, the remaining ones have been shown. Now, you could also ask for an alphanumerical 
character at the end of a line. And you see that again those have g or H or small 
h or 2 are being shown, but the once which control character or a punctuation or a null or a equal 
to sign, etc., they are all being skipped.
You can now look for lines that contain digits 
anywhere in the line. And you could also ask the digit should be at the beginning of the 
line or it should be at the end of the line. So, you could see that when you do not provide any 
anchor, it would match anywhere in the line. So, here the 5 is somewhere in the middle of the 
line, here it is in the beginning of the line. And

in the case you provide an anchor with the hat, it 
means it has to be matched only in the beginning of the line. So, the 5g and 42 are being matched. 
And when you say end of the line as an anchor, then only the 42 line will be shown as matched 
line from the file chartypes dot txt.
Now, there are some control characters that I have 
placed in this file. And we could see whether you could pick the lines that contain the control 
characters. And you see that these are the three lines. So here I have got a tab character and 
here I have got a control c sitting here and here also I have got a control character sitting in the 
front. So, those three lines only have been picked up. Now, you could ask a reverse question saying 
that I do not want any control character.
Now, you could also ask that you would like 
to have only the lines which do not contain

the control character and the minus v option 
would actually give you the opposite effect, where the lines that are not matching are 
shown. And you see that the remaining lines, include the empty line, are being shown.
Now, you could also look for any punctuation, characters to be matched. 
So, let us look at that punct, would match for all punctuation characters. So, 
all those lines which have punctuation characters, colon is one of the punctuation characters, so 
that has been matched nicely. And almost every, line has that except the empty line, so empty 
line has been skipped. So, now I give an anchor, at the beginning of the line. 
So, only those lines at the beginning, contain a punctuation mark will be shown. So, 
we have got a comma there in the first, this, line here so that has been matched. And you could 
also ask for a line that ends with a punctuation, and you see that the ones with have a dot and 
the equal to sign have been matched. So, the

punctuation marks without specifying what those 
marks are can also be used as a character set.
  Now, the case of the character also can be used. 
So, if I want to have matching lines that contain only, that contain lower case characters, so you 
would see that all the lines except the empty line are being shown because they all contain some 
lowercase letters. And I would put an anchor here to show that the lowercase characters should be 
matched at the beginning of the line. And so the hello and l will be getting matched there with 
a capital, with a small h and l and I could give an anchor to indicate that I want the matching 
to be done at the end and the 5g line and the h in the end will be matched because they contain a 
lowercase character at the end of the string.
  You could also look at the same kind of 
analysis using the uppercase characters. So, all those lines which contain any uppercase 
character will be displayed and you see that

we do not have uppercase characters used 
very much, so only three lines are used. So, you have got a capital H here, capital L here, and 
capital C here. So, only three lines have been matched. And I will put an anchor here. 
So, only the ones which have a starting with an uppercase character will be shown. 
So, capital L is getting matched there. And I put a dollar here and remove the hat. So, I get a matching where the ending of the line will be a capital letter, that would be the capital H. So, now you see that you could also specify the matching pattern using a character set that contains the uppercase characters.
The alphanumerical characters, as well as the punctuation marks, are combined together as printable characters, because they can display on the screen, and the control characters are then avoided. So, you can actually use the class, character class called print to indicate that you would like to match only printable characters.

And when you want to display, you would see that every line except the blank line has been matched, because except the blank line every other line has some printable characters or other. Now, you could also ask for a printable character to be at the beginning of the line. And you would see that those lines which have a control character in the beginning have been skipped. Then you could also ask for the printable character to be the end of the line. And you see that some of the lines which contain a control character at the end are skipped. So, here, for example, here there is a control character at the end that has been skipped. And here there is a control character in the beginning and that has been kept because the printable character is matched only in the end and therefore in the beginning can be anything else. Now, there are some characters which give you a blank space like a space character or a tab.

character, so they can also be matched. So, 
these are the whitespace characters. And you   would see that except the empty line, 
all the lines contain some whitespace   or other and that has been matched.
And you could also ask for the anchor   to show that in the beginning or to show that in 
the end. So, you see that there is a blank space   that is available at the beginning or at the end 
of these two lines so they are being shown, but   the remaining lines have the blank space somewhere 
in the middle so those are being picked up.
Now, there is a character case called space 
that can also be used to pick up lines that   contain some whitespace. And you could also ask 
for it to have the beginning with the whitespace   or the end with this whitespace. Now, this is 
not very different from the blank space except   that there are more whitespace characters than 
what we have used. And you can see what type

of a whitespace that is allowed for your input file, and accordingly filter them out using the difference between these two character sets. Now, the counter class called graph is used to match any non-space character. So, you can see that every line except the blank line has some character which is not a space, and therefore that is getting matched here. And you could then have that as an anchor in the front. And you could see that those lines that contain any non-space character are being displayed in the front. So, some of them which contain a control C that has been skipped here, a blank that has been skipped here, so the remaining ones only are there, only the remaining ones are being displayed. And I could also put an anchor at the end, and you would see that the lines that contain a blank space at the end are skipped. So, a non-space character is tolerated at the end of the line. Now, let us say we would like to skip

all the empty lines. So that 
can be done using this option. So, what happens is that the matching lines for 
the two anchor characters side by side would give empty lines and minus we would say that do 
not print that, print whatever does not match the empty line. So, the non-empty lines are getting 
printed out now. And the empty lines if you want to print, there is nothing funny about it just 
an empty line. So, that is not very useful.
But the with the minus v option, you could 
actually remove the empty lines from an input stream and that is quite useful 
in processing where you expect something to be there on the line and you do not want to have 
empty lines coming in. So, you could use this feature of minus v, and then the two anchors 
side by side for the beginning and the end, so that you skip empty lines.
Time to time we are expected to pick up patterns

from a given text file and then analyze them if 
they correspond to certain fields. So, now let's look at an example file and see whether we can 
pick up such patterns using the grep command.
So, here I am providing a random piece of text 
which contains certain patterns. So, as we know, the Aadhar card numbers have about 12 digits. 
So, I am giving that as one example in one line. And then the pin codes usually have six digits 
so that of IIT Madras is a six-digit one, 600036. There is no other location in Chennai that uses 
the same pin code. So, I am giving that as one example. And then the phone numbers for mobile 
phones are 10 digits and without the country code or the zero for a city dialing, the landline 
numbers also are 10 digits for the cities. So, I am giving that as an example.
The roll numbers in IIT for regular students

are following the pattern where the first two 
letters correspond to the department code. And then the two digits correspond to the year 
of joining. And then there is a program code which is in one letter followed by three digits 
corresponding to the roll number within the class. And it is not necessary that we need to give the 
URLs of our websites with the protocol like HTTPS or HTTP, but if you just give a string where there 
is a dot and on either sides of the dot there is a word which is alphanumeric, then that is also a 
valid URL. And you would have seen on your mobile phone that such strings are automatically 
made clickable and they could also be, and they could be interpreted as URLs.
So, now, let us see whether we can pick these numbers from this text using the grep command. 
Now, it would not work if any of these are spanning across two lines, because grep would use 
only one line at a time for processing. But with

that limitation very quickly. We can actually put the grep knowledge that we have earned till now to some good use already. So, first, I would like to pick up the numbers which are having exactly 12 digits long so that the other card could come out of the text. So, I would like to always keep printing the original text so that it is useful for you. So, egrep and then the pattern will be given. The pattern should be such that any digit should be matched and it should have exactly 12 occurrences. So, 12 occurrences should be there. And you see that we have identified the Aadhar card as a matching pattern within the first line. So, no other line is getting displayed. So, that means that there is no confusion there. Now, let us say we would like to pick up the pin code. So, now you see that pin code is six digit, so I want to pick those numbers that contain six digits. Now, you see that there is a

problem here. It is actually picking up numbers, 
which are matching six digits of the Aadhar card, as well as for the phone number. 
Now, how do I pick only the six digits, and not more or not less, and that can be done 
using the word boundary. So, now, let us say that before and after the six digits, there must be 
a word boundary. And now you see that it would actually identify only the line that corresponds 
to the pin code. The phone number and the Aadhar card are not coming because they do not have a 
word boundary after the six digits are matched.
Now, I would like to identify what looks like a 
roll number. Now, you see that the roll number is actually a mix of both alphabetical as 
well as the digits. So, it is a bit painful. Now, alphanumerical characters of that particular 
length can also be tried out. Let us do that.

So, let us try out the alphanumerical characters alnum, and we see that it is actually having a length of eight. So, I put an 8 here. So, any eight-character long alphanumerical string would be matched, and you see that is actually not working out because it is matching given regular words which do not have specific digits. Now, no point even asking for a word boundary because a word boundary would also catch words which are exactly eight characters long, so even that does not work. So, we need some other strategy; that is where the character set that we have discussed would come off use. So, now let us try that out. So, what we would do is that we would use a word boundary, and then the alphabetical characters should occur exactly two times following by that a digit has to come exactly two times, and then an alphabetical character has to come.

and after that, again, a digit has to come exactly three times, after that, there is again a word boundary. So, you see that the pattern when you type it is quite long, but it is quite logical. What we are doing is I am asking for a word boundary, after that, I am asking for an alphabetical character coming exactly twice, and then I am asking for a digit that comes exactly twice, and then again an alphabetical character, only one, and then I have got a digit that comes exactly thrice, after that, again, a word boundary. So, this is a pattern that we are giving and you see that it identifies exactly the roll number and nothing else. So, by carefully using the character sets, we can actually identify the pattern of the string that we are trying to pick up. Now, for the URLs also, we can actually use the same kind of a strategy. What is the strategy? So, you know that the protocols are not necessary

to be used now. Even the names like this also 
are possible. So, what we would do is as follows.   And then an alphanumeric character 
can come any number of times, but not   zero, after that a dot has to come, but 
the dot has to be also escaped. And again,   alphanumerical character can be there and that 
also has to come once or more, but definitely   not zero times and then the word boundary. 
So, you see now that it identifies the two URLs   quite nicely. And so you can see that the usage 
of the word boundary here and then the plus sign,   which is available in the extended regular 
expression engine under egrep, and then the   backslash to escape the meaning of the 
dot so that the dot is matched as it is,   and then again, a word boundary to signal that 
is at the end of a particular word would actually   help you in identifying the URLs quite easily. 
Now, I have used only alphanumerical characters.

Underscore is also an allowed character. So, you 
can actually expand the scope of this matching to include every possible URL that can come in 
a random text that you are not expecting.
Now, time to time, there is a need for us to trim 
a text from the top to the bottom or sideways. So, top to bottom, we have already learned how to 
trim by using the command a head and tail. So, the top end lines of a text file can be picked 
up using the head command, the bottom end lines can be picked up using the tail command. But what 
about trimming the text from the left and right where we can use either the number of 
fields or the number of characters. So, for that, there is a function that is 
available. So, the command is called cut.
So, when you use man cut, you 
would see that it is enabling a horizontal trimming of the lines and that 
is used to remove sections from each of the

files and it is going line by line. And it can 
actually trim using either the characters or fields, but for any field you also need 
to know what is the character used for the delimiter so minus d option also has to be 
there. And the default is tab, but you can provide any other delimiting character.
So, now I have prepared a simple text file using which we can actually start trimming 
to understand the usage of the cut command. So, fields dot txt is a simple three line file 
in which I have got fields like this. So, 1, 2, 3, 4 and then a semicolon, after that hello world 
has a blank, after that there is a comma. So, you can see that you could use the semicolon or 
the comma or the blank space as a field separator, and you can also use the number of characters as a 
way to cut. So, now let us go and inspect the how

the cut command can be done.
Now, it can be used both ways. So, you can actually use cut fields dot txt, and then you can actually give some options how to cut. So, minus c is the number of characters, and 1 to 4 tells you that from 1 to 4 characters from the beginning should be taken out and displayed. So, you can see that the first four characters have been taken out and displayed. Now, you could also say that I want from 5 to 8, and I have got the next four characters coming on the screen.
Now, this does not pay attention to what was the field. It just goes as a series of characters and picks up the character sequence that you want. So, 1 to 4 or 5 to 8 or whatever. And if you do not specify the initial part, it will actually assume that you are going to have the, from the beginning.

okay. And if you do not specify the ending, it 
will also assume that you want it till the end. So, the option is there for the hyphen either the 
left side of the hyphen or the right hand side of the hyphen can be provided. If you provide 
both then it will be indicating as a range.
Now, let us use a delimiter to pick up some of the 
fields. So, let us use it as a piped command. So, cut can take the input from a pipe. So, we will do 
that now. And let us use the delimiter as a space. So, now you will see that if I use the delimiter 
as a space, then this will be the first field and this is the second field. Now, 
I want to print only the first field. 
Now, you see that it comes out quite nicely. 
And the length of each of these fields need

not be same. So which means that when the 
number of characters is not what is uniform, but the way the characters are 
separated is what is uniform, then using space as a delimiter you can pick 
the first field. You can also pick the second field and you see that the second half of 
the text is coming up onto the screen. Now, if you want both, you can also 
get it. So, then you have got both the entire lines coming up, so that is actually not 
meaningful in this case. But in some situations where there are more than two fields, then you 
can actually take part of those fields also. Now, let us use the semicolon as a delimiter. And 
you would see that when I use semicolon, then the first field is this number, the second field is a 
string that contains a comma also in between. So, let us check that out. And you see that only the 
numbers are coming when I asked the first field.

And when I asked the second field, I am getting the string, and then the comma. So, this is quite nice. You can change the delimiter and cut fields as per the delimiter, and pick up only those fields that you would like to have. So, similarly, we can actually also use comma as a delimiter. And you see that when I have comma as a delimiter, the second field will contain the line number. So, I would like to take the second field, and you see only the line numbers are being displayed here, because after the comma, this is the second field, this is the second field. And if you want the first field before the comma, the rest of it actually is coming onto the screen, which actually does not have the same length for each of the lines. So, the comma, the semicolon, the space or the tab, all of these can be used as field

separators. So, look at the input stream once 
with your own eyes to see what is the delimiter, that was used by the program which is 
throwing that particular stream at you. And then once you understand, you can go ahead and 
use the cut command to pick the fields that you want without having to tell the original developer 
to provide them as you like, so you can just cut and take what you need from the output stream.
Now, you can also combine the streams by piping them multiple times. Let us say I 
would like to pick only the string in between the semicolon and the comma, so what I will 
do is I will use semicolon as the first field separator and then comma as a second 
time. So, let us look at that now.
So, I want to pick only this part. So what 
I would do is, I would say cat fields, then cut the minus delimiter, I would first use 
this, and take the first field, and then I would

use comma as the delimiter, and again take the 
first field. So, second field for the semicolon, as delimiter because I want the second half of 
it, and then after that I will separate using the comma, and then I will take the first half of it. 
So, that is what I am meaning by this command. And you see that I am getting the string that is 
exactly between the semicolon and the comma.
So, now, this is actually not bothering about 
anything else. So, you can also achieve this to some extent using grep also. But in this 
case, we are looking at the text in a slightly different manner. We are looking at as a set of 
fields that are separated and picking the fields, which is slightly different way of looking 
at than a grep command. But we can also try to see whether the grep would work to give the 
same result. So, let us try that out now.

And you see that the grep command was able to 
match exactly the part that you would like to match. Now, but you have not cut it off. So, 
you have not yet trimmed the rest of it off. For the trimming part, you can 
pass on the output of the grep to said command which we would learn in the 
future classes, but for now you can see that you can also identify the string that you 
want by using pattern matching quite well. But the cut command is very useful because you can 
actually remove that portion and then go on to look at the output in a way that you would like 
to parse it with any other tool later on.
Now, you could actually combine the concept 
of cutting the text from left and right using the cut command and then the head and tail to 
actually cut the lines from the top and the bottom to pick a particular entry from the input 
string. So, let us try that out and I want

to pick up only the welcome command line as 
the text. So, what I would do is, I will cut   it from the left and right first and after that 
top and bottom. So, let us look at that now.
  So, here I am getting the three lines. Now, 
I would like to only have the second line.   So, I would do like this, head minus n and then 
two lines. The top two lines are being picked up.   And in the top two lines, I want the last line. So, 
I would pick the last line coming out. And now you   see that I have picked up exactly the string 
welcome command line coming on to the screen   and I have used many, many commands here.
So, I have catenated the contents of the file   fields dot text and I have cut it using the field 
separator semicolon and picked up the second field   and after that I have again cut it using the comma 
as a field separator and pick the first field

and after that, I have picked up the top two lines. And then from those top two lines, I have taken the bottom-most line and then displayed it. And we have got the string that we wanted. So, you can actually combine the commands in various ways to get the output that you are interested in. And it is up to our imagination to find the easiest way. And of course, the most elegant way would be when we are experts in all these commands for which a lot of practice is required. So, keep practicing various of these commands and be thorough with regular expressions so that you make the best use of the tools that are available in the Linux command-line environment.

Pattern Matching - Part 02

Pattern Matching - Part 02