Practice with regular expressions

This is not a graded assignment but rather optional preparation for assignment 1. Please feel free to ask each other, David, or Emily for assistance with any aspect of this assignment, including how to get your computer set up to do it.

In this exercise, you will use grep (generalized regular expression printer) to get practice with writing regular expressions in such a way that the computer will give you feedback on how you're doing. You can find grep on dante (which is a unix system), or, on a mac with OSX, by using the terminal program.

Note: In the following, I've used "\n" to indicate that you should type return.

Getting started, on Dante:

Log on to Dante using ssh from one of the PCs in the lab (LLC 112), or elsewhere. On the PCs in the lab, ssh assumes you want to use it for ftp, which you don't in this case. Instead, click the button called "quick connection" (EB/DG: double check this).
At the first screen once you're on Dante, type S. This puts you into a Unix shell.
Type "which grep\n". It should respond with something like "/bin/grep". This means that there is a program called grep on the system, and the computer knows where to find it.
Grep takes two arguments, a regular expression to search for and a file to search in. To do this exercise, you'll need to make a file to do the searching, with this content. To make the file, using emacs:
- Cut and paste the text found by following this link: grepfile into the emacs window.
- Save the file by typing ctrl-x s.
- Quit emacs by typing ctrl-x ctrl-c. Your Dante window should now be back to the state where it gives you a prompt like "dante05%".
- Type "ls\n" to verify that you have created the file, and "cat grepfile\n" to verify its contents.
Continue with the exercises below.

Getting started, on a Mac

Run the terminal program.
Type "which grep\n". It should respond with something like "/usr/bin/grep". This means that there is a program called grep on the system, and the computer knows where to find it.
Change directories within the terminal program to your documents directory by typing: "cd ~/Documents".
Grep takes two arguments, a regular expression to search for and a file to search in. To do this exercise, you'll need to make a file to do the searching, with this content. You can make this file with textedit or emacs (see the instructions under 'getting started, on Dante' above). You can also use MS Word, but in that case be sure to save your file as text only. Be sure to save the file in the Documents directory in your home directory.
Continue with the exercises below.

Exercises

At your unix prompt, type "grep 'dog'\n". Observe that nothing happens. This is because grep is waiting for the other half of its input. Type ctrl-c. Observe that you get your prompt back.

At your unix prompt, type "grep 'dog' grepfile\n". Observe what happens.

Before proceeding, some notes on grep and regular expressions:

It's a good idea to enclose your regular expression in single quotes ('), because otherwise the shell will try to interpret some of the special characters, notably *.
Some of the regular expression syntax is different for grep than for perl. Refer to the table on page 831 of the textbook. Note that grep doesn't have + or ? as part of its regular expression syntax.

Now work out a regular expression to use that will make grep return only the line containing just the word dog. Test it by typing "grep 'xxx' grepfile\n', with your regular expression instead of xxx.

Now work out a regular expression that will make grep return only the lines "dog" and "doog".

Work out a different regular expression that will make grep return only the lines "dog" and "doog".

Work out a regular expression that will make grep return only the line "dog dog".

Work out another one.

And another one. (If you haven't used $ and $ yet, try.)

Do the exercise 2.1, parts a-f, on paper. Use either Perl or grep regular expression syntax.

Consider these sample answers to the exercise. How are they different from your answers?
Look at this version of the sample answers, where I've given prose descriptions of the regular expressions. Make sure you understand how the prose corresponds to the regular expressions.
Using emacs (or, on a mac, your favorite text editor) as described above, make a file to test the answers to 2.1. Be sure to include strings that should match and similar strings that shouldn't.
Using grep and that file, test the regular expressions. Be sure to use the ones written in grep syntax.
- Consider testing any regular expressions that you wrote differently from the sample answers to see how they behave differently.
- You may need to add beginning and end of line anchors (^ and $) to make sure you're not just matching substrings. (This is again the difference between regular expressions as definitions of regular languages, and regular expressions used in search.)