# Reading data from files

Summary: We begin our consideration of ways to store data in files and how to work with data once they are stored in files. Along the way, we consider some of the complexities of representing data.

## Introduction: Storing data

In our recent exploration of tables we found that it is possible to work with tables of data when they are represented as lists of lists. And that makes sense if we’re working in Scheme. However, people work with data in many languages, so we should not store it in Scheme format. Fortunately, there are a variety of formats that computer scientists use to store data. We will explore a few such formats and the procedures we use to work with data in those formats.

## Comma-separated values

The most common file format for tables of data are comma-separated value files, most typically referred to by the initials “CSV” or “csv”. Each line of a csv file contains one row of a table, with the entries separated by commas (hence the name). In general, the assumption in csv files is that something that looks like a number is a number and everything else is a string. For example, here is an entry in our “state capitals” file for Des Moines.

Iowa,Des Moines,41.590939,-93.620866


At first glance, the basic rules seem simple and straightforward: (1) values are separated by commas, (2) numbers represent numbers, (3) everything else is a string. But when we start to probe the details, things get a bit more complicated.

What if we want to include a comma in one of the entries, such as “The Roswell, New Mexico Crash” from our UFO table? The designers of csv sensibly decided that we should be able to put strings that include commas in quotation marks. Hence, we might write the following.

1947,"The Roswell, New Mexico Crash","July, 1947",USA,Yes,Yes,Yes,No


Quotation marks are also useful when we want to represent something that looks like a number, but should be treated verbatim. For example, zip codes look like numbers. However, we typically want the leading zeros, such as 02158 rather than 2158. In representing our data that includes zip codes, we would therefore put each zip code in quotation marks.

"02158",42.385096,-71.208399,"Newton","MA","Middlesex"


There are a few other subtleties, such as how you represent a string that includes both a comma and a double-quotation mark. But those are rare enough that we’ll leave them as an appendix.

## Reading from csv files

The csc151 package provides one primary procedure for reading from csv files, read-csv-file. It takes as its parameter a string that describes the location of the file on the filesystem. For example, we might write

> (read-csv-file "/home/rebelsky/share/csv/courses.csv")
'(("Rebelsky" "2017F" "CSC-151" 1 4)
("Klinge" "2017F" "CSC-151" 2 4)
("Rebelsky" "2017F" "CSC-151" 3 4)
...
("Rebelsky" "2018S" "CSC-321" 1 2)
("Rebelsky" "2018S" "CSC-322" 1 2)
("Klinge" "2017F" "CSC-341" 1 4))


As this example suggests, we are likely to get a lot of data. So we almost always use read-csv-file in conjunction with a definition.

> (take courses 3)
'(("Rebelsky" "2017F" "CSC-151" 1 4)
("Klinge" "2017F" "CSC-151" 2 4)
("Rebelsky" "2017F" "CSC-151" 3 4))


Once we’ve read a csv file, we can treat it just as we’ve been treating any other list of lists.

You may also find files that are much like csv files but that use a different character as a separator. The most common separator other than comma is a tab. To read such files, you use the awkwardly named read-csv-file-w/sep, which takes two parameters: the file name and the character used for separating entries.

## Naming files

Racket is surprisingly clueless about finding files. We might say “It’s right there.” But there is not clear to the computer. Hence, we will generally identify files by their full path name. As the example above suggested, when we are working on a Unix/Linux system, we most typically use a path name of the form "/home/username/path/to/file". For example, if us-zip-codes.csv is sitting on your desktop, you would write "/home/username/Desktkop/us-zip-codes.csv".

If you are working on a Mac, you start the path with /Users rather than /home, as in "/Users/username/Desktkop/us-zip-codes.csv".

Windows path names are much more complicated. Read the Racket documentation if you want to deal with files on Windows.

## Working with plain text files

While csv files are the most common form of tabular data, there are certainly other kinds of data we might want to work with. One of the simplest is what we would call a “plain text file”, a file that just contains words. Plain text files often have a .txt suffix, although you’ll also find variants.

Computer scientists and data scientists work with text files in a variety of ways. They might, for example, search for particular words or attempt to rewrite the text in a file into a new form or a new language. They might look for some statistical properties of the text to try to gain some insight. We might (and will) also consider similar issues.

The csc151 package provides three basic operations for working with text files: file->chars, which reads the contents of a text file and presents the contents as a list of characters; file->words, which reads the contents of a text file and presents the contents as a list of strings, each of which represents one “word” in the file (using a simple metric for “word”); and file->lines, which reads the contents of a text file and presents the contents as a list of strings, each of which represents one line of the input file.

Suppose we had the previous paragraph in a file. Here’s what we might get reading it each way.

> (take (file->chars "/home/rebelsky/share/text/csc151-textfile-procs.txt") 20)
'(#\T #\h #\e #\space #\ #\c #\s #\c #\1 #\5 #\1 #\ #\space #\p #\a #\c #\k #\a #\g #\e)
> (take (file->words "/home/rebelsky/share/text/csc151-textfile-procs.txt") 10)
'("The" "csc151" "package" "provides" "three" "basic" "operations" "for" "working" "with")
> (take (file->lines "/home/rebelsky/share/text/csc151-textfile-procs.txt") 5)
'("The csc151 package provides three basic operations for working with"
"text files: file->chars, which reads the contents of a text file and"
"presents the contents as a list of characters; file->words, which"
"reads the contents of a text file and presents the contents as a list"
"of strings, each of which represents one \"word\" in the file (using a")


We can then process those lists in different ways. We will begin in the lab by exploring tallies.

## Self checks

### Check 1: Your own csv file

a. Create a csv file that represents your schedule for the semester. For example, a line of your file might read

CSC-151,01,MWF,2:30,3:50,Professor Rosseforp


b. Confirm that you can read the file using read-csv-file.

### Check 2: Reading text

a. What do you expect to happen if you call file->chars, file->words, and file->lines on the csv file you created in the first self check?