r4ds/import.Rmd

630 lines
25 KiB
Plaintext
Raw Normal View History

2015-12-12 02:34:20 +08:00
# Data import
2016-07-07 02:59:50 +08:00
## Introduction
2016-01-27 21:33:35 +08:00
2016-07-12 04:38:39 +08:00
Working with existing data is a great way to learn the tools of data science, but at some point you want to stop learning and start working with your own data. In this chapter, you'll learn how to use the readr package for reading plain-text rectangular files into R.
This chapter will only scratch surface of data import, many of the principles will translate to the other forms of data import. The chapter concludes with a few pointers to packages that you might find useful.
2016-01-27 21:33:35 +08:00
2016-07-07 02:59:50 +08:00
### Prerequisites
2015-09-23 02:35:39 +08:00
2016-07-07 02:59:50 +08:00
In this chapter, you'll learn how to load flat files in R with the readr package:
2015-09-23 02:35:39 +08:00
2016-07-07 02:59:50 +08:00
```{r setup}
library(readr)
```
2015-09-23 02:35:39 +08:00
2016-07-12 04:38:39 +08:00
## Getting started
2015-09-23 02:35:39 +08:00
Most of readr's functions are concerned with turning flat files into data frames:
2016-01-27 21:33:35 +08:00
* `read_csv()` reads comma delimited files, `read_csv2()` reads semi-colon
2015-09-23 02:35:39 +08:00
separated files (common in countries where `,` is used as the decimal place),
`read_tsv()` reads tab delimited files, and `read_delim()` reads in files
2016-07-07 02:59:50 +08:00
with any delimiter.
2015-09-23 02:35:39 +08:00
* `read_fwf()` reads fixed width files. You can specify fields either by their
2016-01-27 21:33:35 +08:00
widths with `fwf_widths()` or their position with `fwf_positions()`.
2015-09-23 02:35:39 +08:00
`read_table()` reads a common variation of fixed width files where columns
are separated by white space.
* `read_log()` reads Apache style logs. (But also check out
2016-01-27 21:33:35 +08:00
[webreadr](https://github.com/Ironholds/webreadr) which is built on top
2015-09-23 02:35:39 +08:00
of `read_log()`, but provides many more helpful tools.)
2016-07-07 02:59:50 +08:00
These functions all have similar syntax: once you've mastered one, you can use the others with ease. For the rest of this chapter we'll focus on `read_csv()`. If you understand how to use this function, it will be straightforward to apply your knowledge to all the other functions in readr.
2015-09-23 02:35:39 +08:00
2016-07-07 02:59:50 +08:00
The first argument to `read_csv()` is the most important: it's the path to the file to read.
2015-09-23 02:35:39 +08:00
2016-07-07 02:59:50 +08:00
```{r}
heights <- read_csv("data/heights.csv")
```
2015-09-23 02:35:39 +08:00
2016-07-12 04:38:39 +08:00
When you run `read_csv()` it prints how out a column specification that gives the name and type of each column. That's an important part of readr, which we'll come back to in [[parsing a file]].
2015-09-23 02:35:39 +08:00
2016-07-08 00:17:11 +08:00
You can also supply an inline csv file. This is useful for experimenting and creating reproducible examples:
2015-09-23 02:35:39 +08:00
2016-07-07 02:59:50 +08:00
```{r}
read_csv("a,b,c
1,2,3
4,5,6")
```
2016-01-27 21:33:35 +08:00
2016-07-08 00:17:11 +08:00
Notice that `read_csv()` uses the first line of the data for column headings. This is a very common convention. There are two cases where you might want tweak this behaviour:
2015-09-23 02:35:39 +08:00
2016-07-07 02:59:50 +08:00
1. Sometimes there are a few lines of metadata at the top of the file. You can
use `skip = n` to skip the first `n` lines; or use `comment = "#"` to drop
all lines that start with a comment character.
```{r}
2016-07-08 00:17:11 +08:00
read_csv("The first line of metadata
The second line of metadata
2016-07-07 02:59:50 +08:00
x,y,z
2016-07-08 00:17:11 +08:00
1,2,3", skip = 2)
2016-07-07 02:59:50 +08:00
read_csv("# A comment I want to skip
x,y,z
1,2,3", comment = "#")
```
1. The data might not have column names. You can use `col_names = FALSE` to
tell `read_csv()` not to treat the first row as headings, and instead
label them sequentially from `X1` to `Xn`:
```{r}
read_csv("1,2,3\n4,5,6", col_names = FALSE)
```
Alternatively you can pass `col_names` a character vector which will be
used as the column names:
```{r}
read_csv("1,2,3\n4,5,6", col_names = c("x", "y", "z"))
```
2015-09-23 02:35:39 +08:00
2016-07-12 04:38:39 +08:00
Another option that commonly needs tweaking is `na`: this specifies the value (or values) that are used to represent missing values in your data:
```{r}
read_csv("a,b,c\n1,2,.", na = ".")
```
This is all you need to know to read ~50% of csv files that you'll encounter in practice. You can also easily adapt what you've learned to read tab separated files with `read_tsv()` and fixed width files with `read_fwf()`. To read in more challenging files, you'll need to learn more about how readr parses each individual column, turning a character vector into the most appropriate type.
2016-07-07 02:59:50 +08:00
### Compared to base R
2015-09-23 02:35:39 +08:00
2016-07-08 00:17:11 +08:00
If you've used R before, you might wonder why we're not using `read.csv()`. There are a few good reasons to favour readr functions over the base equivalents:
2016-01-27 21:33:35 +08:00
2016-07-08 00:17:11 +08:00
* They are typically much faster (~10x) than their base equivalents.
2016-07-12 04:38:39 +08:00
Long running jobs have a progress bar, so you can see what's happening.
Note that if you're looking for raw speed, try `data.table::fread()`. It
doesn't fit quite so well into the tidyverse, but it can be quite a bit
faster.
2016-01-27 21:33:35 +08:00
2016-07-08 00:17:11 +08:00
* They produce tibbles, and they don't convert character vectors to factors,
2016-07-12 04:38:39 +08:00
produce row names, or munge the column names. These are common sources of
frustration with the base R functions.
2016-01-27 21:33:35 +08:00
2016-07-08 00:17:11 +08:00
* They are more reproducible. Base R functions inherit some behaviour from
2016-07-12 04:38:39 +08:00
your operating system and environment variables, so import code that works
on your computer might not work on someone else's.
2016-01-27 21:33:35 +08:00
2016-07-09 05:23:19 +08:00
### Exericses
2016-07-12 04:38:39 +08:00
1. What function would you use to read a file where fields were separated with
"|"?
2016-07-09 05:23:19 +08:00
1. Apart from `file`, `skip`, and `comment`, what other arguments do
`read_csv()` and `read_tsv()` have in common?
2016-07-12 04:38:39 +08:00
1. What is the most important argument to `read_fwf()` that we haven't already
discussed?
2016-07-09 05:23:19 +08:00
1. Some times strings in a csv file contain commas. To prevent them from
causing problems they need to be surrounded by a quoting character, like
`"` or `'`. By convention, `read_csv()` assumes that the quoting
character will be `"`, and if you want to change it you'll need to
use `read_delim()` instead. What arguments do you need to specify
to read the following text into a data frame?
```{r}
"x,y\n1,'a,b'"
```
2016-07-12 04:38:39 +08:00
1. Identify what is wrong with each of the following inline csvs.
What happens when you run the code?
```{r, eval = FALSE}
read_csv("a,b\n1,2,3\n4,5,6")
read_csv("a,b,c\n1,2\n1,2,3,4")
read_csv("a,b\n\"1")
read_csv("a,b\n1,2\na,b")
read_csv("a;b\n1;3")
```
2016-07-09 05:23:19 +08:00
2016-07-08 00:17:11 +08:00
## Parsing a vector
2015-09-23 02:35:39 +08:00
2016-07-12 04:38:39 +08:00
Before we get into the details of how readr reads files from disk, we're need to take a little detour to talk about the `parse_*()` functions. These functions take a character vector and return a more specialised vector like a logical, integer, or date:
2015-09-23 02:35:39 +08:00
2016-07-08 00:17:11 +08:00
```{r}
str(parse_logical(c("TRUE", "FALSE", "NA")))
str(parse_integer(c("1", "2", "3")))
str(parse_date(c("2010-01-01", "1979-10-14")))
```
2015-09-23 02:35:39 +08:00
2016-07-08 00:17:11 +08:00
These functions are useful in their own right, but are also an important building block for readr. Once you've learned how the individual parsers work in this section, we'll circle back and see how they fit together to parse a complete file in the next section.
2016-07-12 04:38:39 +08:00
Like all functions in the tidyverse, the `parse_*()` functions are uniform: the first argument is a character vector to parse, and the `na` argument specifies which strings should be treated as missing:
2015-09-23 02:35:39 +08:00
2016-07-07 02:59:50 +08:00
```{r}
2016-07-08 00:17:11 +08:00
parse_integer(c("1", "231", ".", "456"), na = ".")
2016-07-07 02:59:50 +08:00
```
2015-09-23 02:35:39 +08:00
2016-07-09 05:23:19 +08:00
If parsing fails, you'll get a warning:
2015-09-23 21:58:16 +08:00
2016-07-07 02:59:50 +08:00
```{r}
2016-07-08 00:17:11 +08:00
x <- parse_integer(c("123", "345", "abc", "123.45"))
2016-07-09 05:23:19 +08:00
```
And the failures will be missing in the output:
```{r}
x
```
2016-07-12 04:38:39 +08:00
If there are many parsing failures, you'll need to use `problems()` to get the complete set. This returns a tibble which you can then explore with dplyr.
2016-07-09 05:23:19 +08:00
```{r}
2016-07-08 00:17:11 +08:00
problems(x)
2015-09-23 21:58:16 +08:00
```
2016-07-12 04:38:39 +08:00
Using parsers is mostly a matter of understanding what's available and how they deal with different types of input. There are eight particularly important parsers:
2016-07-08 00:17:11 +08:00
1. `parse_logical()` and `parse_integer()` parse logicals and integers
2016-07-12 04:38:39 +08:00
respectively. There's basically nothing that can go wrong with these
parsers so I won't describe them here further.
2016-07-08 00:17:11 +08:00
1. `parse_double()` is a strict numeric parser, and `parse_number()`
is a flexible numeric parser. These are more complicated than you might
expect because different parts of the world write numbers in different
ways.
1. `parse_character()` seems so simple that it shouldn't be necessary. But
2016-07-09 05:23:19 +08:00
one complication makes it quite important: character encodings.
2016-07-08 00:17:11 +08:00
1. `parse_datetime()`, `parse_date()`, and `parse_time()` allow you to
parse various date & time specifications. These are the most complicated
because there are so many different ways of writing dates.
The following sections describe the parsers in more detail.
2015-09-23 21:58:16 +08:00
2016-07-07 02:59:50 +08:00
### Numbers
2015-09-23 21:58:16 +08:00
2016-07-12 04:38:39 +08:00
It seems like it should be straightforward to parse a number, but three problems make it tricky:
2016-01-27 21:33:35 +08:00
2016-07-07 02:59:50 +08:00
1. People write numbers differently in different parts of the world.
2016-07-09 05:23:19 +08:00
Some countries use `.` in between the integer and fractional parts of
2016-07-12 04:38:39 +08:00
a real number, while others use `,`.
2016-07-07 02:59:50 +08:00
2016-07-08 00:17:11 +08:00
1. Numbers are often surrounded by other characters that provide some
context, like "$1000" or "10%".
2016-01-27 21:33:35 +08:00
2016-07-08 00:17:11 +08:00
1. Numbers often contain "grouping" characters to make them easier to read,
2016-07-12 04:38:39 +08:00
like "1,000,000". The characters that are used to group numbers into chunks
differ around the world.
2016-07-08 00:17:11 +08:00
2016-07-12 04:38:39 +08:00
To address the first problem, readr has the notion of a "locale", an object that specifies parsing options that differ from place to place. When parsing numbers, the most important option is the character you use for the decimal mark:
2015-09-23 21:58:16 +08:00
2016-07-07 02:59:50 +08:00
```{r}
parse_double("1.23")
parse_double("1,23", locale = locale(decimal_mark = ","))
```
2015-09-23 21:58:16 +08:00
2016-07-12 04:38:39 +08:00
readr's default locale is US-centric, because generally R is US-centric (i.e. the documentation of base R is written in American English). An alternative approach would be to try and guess the defaults from your operating system. This is hard to do well, but more importantly makes your code fragile: even if it works on your computer, it might fail when you email it to a colleague in another country.
2015-09-23 21:58:16 +08:00
2016-07-12 04:38:39 +08:00
`parse_number()` addresses the second problem: it ignores non-numeric characters before and after the number. This is particularly useful for currencies and percentages, but also works to extract numbers embedded in text.
2016-07-07 02:59:50 +08:00
```{r}
parse_number("$100")
parse_number("20%")
2016-07-08 00:17:11 +08:00
parse_number("It cost $123.45")
2015-09-23 21:58:16 +08:00
```
2016-07-12 04:38:39 +08:00
The final problem is addressed by the combination of `parse_number()` and the locale as `parse_number()` will ignore the "grouping mark":
2015-09-23 21:58:16 +08:00
2016-07-07 02:59:50 +08:00
```{r}
2016-07-12 04:38:39 +08:00
parse_number("$123,456,789")
parse_number("123.456.789", locale = locale(grouping_mark = "."))
parse_number("123'456'789", locale = locale(grouping_mark = "'"))
2016-07-07 02:59:50 +08:00
```
2015-09-23 21:58:16 +08:00
2016-07-07 02:59:50 +08:00
### Character
2016-07-08 00:17:11 +08:00
It seems like `parse_character()` should be really simple - it could just return its input. Unfortunately life isn't so simple, as there are multiple ways to represent the same string. To understand what's going on, we need to dive into the details of how computers represent strings. In R, we can get at the underlying binary representation of a string using `charToRaw()`:
```{r}
charToRaw("Hadley")
```
2016-07-12 04:38:39 +08:00
Each hexadecimal number represents a byte of information: `48` is H, `61` is a, and so on. The mapping from hexadecimal number to character is called the encoding, and in this case the encoding is called ASCII. ASCII does a great job of representing English characters, because it's the __American__ Standard Code for Information Interchange.
2016-07-08 00:17:11 +08:00
2016-07-12 04:38:39 +08:00
Things get more complicated for languages other than English. In the early days of computing there were many competing standards for encoding non-English characters, and to correct interpret a string you need to know both the the encoding and the hexadecimal values. For example, two common encodings are Latin1 (aka ISO-8859-1, used for Western European languages) and Latin2 (aka ISO-8859-2, used for Eastern European languages). In Latin1, the byte `b1` is "±", but in Latin2, it's "ą"! Fortunately, today there is one standard that is supported almost everywhere: UTF-8. UTF-8 can encode just about every character used by humans today, as well as many extra symbols (like emoji!).
2016-07-08 00:17:11 +08:00
2016-07-12 04:38:39 +08:00
readr uses UTF-8 everywhere: it assumes your data is UTF-8 encoded when you read it, and always uses it when writing. This is a good default, but will fail for data produced by older systems that don't understand UTF-8. If this happens to you, your strings will look weird when print them. Sometimes you might get complete gibberish, or sometimes just one or two characters might be messed up:
2015-09-23 21:58:16 +08:00
```{r}
2016-07-08 00:17:11 +08:00
x1 <- "\x82\xb1\x82\xf1\x82\xc9\x82\xbf\x82\xcd"
x2 <- "El Ni\xf1o was particularly bad this year"
x1
x2
2015-09-23 21:58:16 +08:00
```
2015-09-23 02:35:39 +08:00
2016-07-08 00:17:11 +08:00
To fix the problem you need to specify the encoding in `parse_character()`:
2016-07-07 02:59:50 +08:00
2016-07-08 00:17:11 +08:00
```{r}
parse_character(x1, locale = locale(encoding = "Shift-JIS"))
parse_character(x2, locale = locale(encoding = "Latin1"))
```
2016-07-07 02:59:50 +08:00
2016-07-12 04:38:39 +08:00
How do you find the correct encoding? If you're lucky, it'll be included somewhere in the data documentation. But that's rarely the case, so readr provides `guess_encoding()` to help you figure it out. It's not foolproof, and it works better when you have lots of text (unlike here), but it's a reasonable place to start. Even then you may need to try a couple of different encodings before you get the right once.
2015-09-23 02:35:39 +08:00
2015-09-23 21:58:16 +08:00
```{r}
2016-07-08 00:17:11 +08:00
guess_encoding(charToRaw(x1))
guess_encoding(charToRaw(x2))
2015-09-23 21:58:16 +08:00
```
2015-09-23 02:35:39 +08:00
2016-07-08 00:17:11 +08:00
The first argument to `guess_encoding()` can either be a path to a file, or, as in this case, a raw vector (useful if the strings are already in R).
2016-07-07 02:59:50 +08:00
2016-07-12 04:38:39 +08:00
Encodings are a rich and complex topic, and I've only scratched the surface here. If you'd like to learn more I'd recommend reading the detailed explanation at <http://kunststube.net/encoding/>.
2016-07-09 05:23:19 +08:00
2016-07-07 02:59:50 +08:00
### Dates, date times, and times
2015-09-23 02:35:39 +08:00
2016-07-12 04:38:39 +08:00
You pick between three parsers depending on whether you want a date (the number of days since 1970-01-01), a date time (the number of seconds since midnight 1970-01-01), or a time (the number of seconds since midnight):
2015-09-23 02:35:39 +08:00
2016-07-10 20:56:06 +08:00
* `parse_datetime()` expects an ISO8601 date time. ISO8691 is an
international standard in which the components of a date are
organised from biggest to smallest: year, month, day, hour, minute,
2016-07-12 04:38:39 +08:00
second.
2016-07-09 05:23:19 +08:00
```{r}
parse_datetime("2010-10-01T2010")
# If time is omitted, it will be set to midnight
parse_datetime("20101010")
```
2016-07-10 20:56:06 +08:00
This is the most important date/time standard, and if you work with
dates and times frequently, I recommend reading
<https://en.wikipedia.org/wiki/ISO_8601>
* `parse_date()` expects a year, an optional separator, a month,
an optional separator, and then a day:
2016-07-09 05:23:19 +08:00
```{r}
parse_date("2010-10-01")
```
2015-09-23 02:35:39 +08:00
2016-07-10 20:56:06 +08:00
* `parse_time()` expects an hour, an optional colon, a minute,
an optional colon, optional seconds, and optional am/pm specifier:
2016-07-09 05:23:19 +08:00
```{r}
library(hms)
parse_time("20:10:01")
```
2016-07-10 20:56:06 +08:00
Base R doesn't have a great built in class for time data, so we use
the one provided in the hms package.
2015-09-23 02:35:39 +08:00
2016-07-10 20:56:06 +08:00
If these defaults don't work for your data you can supply your own datetime formats, built up of the following pieces:
2015-09-23 02:35:39 +08:00
2016-07-09 05:23:19 +08:00
Year
2016-07-12 04:38:39 +08:00
: `%Y` (4 digits).
: `%y` (2 digits); 00-69 -> 2000-2069, 70-99 -> 1970-1999.
2015-09-23 02:35:39 +08:00
2016-07-09 05:23:19 +08:00
Month
2016-07-12 04:38:39 +08:00
: `%m` (2 digits).
: `%b` (abbreviated name, like "Jan").
2016-07-09 05:23:19 +08:00
: `%B` (full name, "January").
2015-09-23 02:35:39 +08:00
2016-07-09 05:23:19 +08:00
Day
2016-07-12 04:38:39 +08:00
: `%d` (2 digits).
: `%e` (optional leading space).
2015-09-23 02:35:39 +08:00
2016-07-09 05:23:19 +08:00
Time
: `%H` 0-24 hour.
: `%I` 1-12, must be used with `%p`.
: `%p` AM/PM indicator.
: `%M` minutes.
: `%S` integer seconds.
: `%OS` real seconds.
: `%Z` Time zone (as name, e.g. `America/Chicago`). Beware abbreviations:
if you're American, note that "EST" is a Canadian time zone that does not
2016-07-12 04:38:39 +08:00
have daylight savings time. It is \emph{not} Eastern Standard Time!
2016-07-09 05:23:19 +08:00
: `%z` (as offset from UTC, e.g. `+0800`).
2015-09-23 02:35:39 +08:00
2016-07-12 04:38:39 +08:00
Non-digits
: `%.` skips one non-digit character.
2016-07-09 05:23:19 +08:00
: `%*` skips any number of non-digits.
2015-09-23 02:35:39 +08:00
The best way to figure out the correct string is to create a few examples in a character vector, and test with one of the parsing functions. For example:
```{r}
parse_date("01/02/15", "%m/%d/%y")
parse_date("01/02/15", "%d/%m/%y")
parse_date("01/02/15", "%y/%m/%d")
```
2016-07-09 05:23:19 +08:00
If you're using `%b` or `%B` with non-English month names, you'll need to set the `lang` argument to `locale()`. See the list of built-in languages in `date_names_langs()`, or if your language is not already included, create your own with `date_names()`.
2015-09-23 02:35:39 +08:00
2016-07-07 02:59:50 +08:00
```{r}
2016-07-09 05:23:19 +08:00
parse_date("1 janvier 2015", "%d %B %Y", locale = locale("fr"))
2016-07-07 02:59:50 +08:00
```
2016-07-10 20:56:06 +08:00
2016-07-09 05:23:19 +08:00
### Exercises
2016-07-12 04:38:39 +08:00
1. What are the most important arguments to `locale()`?
1. I didn't discuss the `date_format` and `time_format` options to
`locale()`. What do they do? Construct an example that shows when
they might be useful.
1. If you live outside the US, create a new locale object that encapsulates
the settings for the types of file you read most commonly.
2016-07-10 20:56:06 +08:00
1. What's the difference between `read_csv()` and `read_csv2()`?
2016-07-09 05:23:19 +08:00
1. What are the most common encodings used in Europe? What are the
2016-07-12 04:38:39 +08:00
most common encodings used in Asia? Do some googling to find out.
2015-09-23 21:58:16 +08:00
2016-07-10 20:56:06 +08:00
1. Generate the correct format string to parse each of the following
dates and times:
```{r}
d1 <- "January 1, 2010"
d2 <- "2015-Mar-07"
d3 <- "06-Jun-2017"
d4 <- "August 19 (2015)"
d5 <- "12/30/14" # Dec 12, 2014
t1 <- "1705"
t2 <- "11:15:10.12 PM"
```
2016-07-08 00:17:11 +08:00
## Parsing a file
2016-07-10 22:19:56 +08:00
Now that you've learned how to parse an individual vector, it's time to turn back and explore how readr parses a file. There are two new things that you'll learn about in this section:
2015-09-23 21:58:16 +08:00
2016-07-12 04:38:39 +08:00
1. How readr automatically guesses the type of each column.
1. How to override the default specification.
2016-07-08 00:17:11 +08:00
2016-07-10 22:19:56 +08:00
### Strategy
2016-07-08 00:17:11 +08:00
2016-07-12 04:38:39 +08:00
readr uses a heuristic to figure out the type of each column: it reads the first 1000 rows and uses some (moderately conservative) heuristics to figure out the type of each column. You can emulate this process with a character vector using `guess_parser()`, which returns readr's best guess, and `parse_guess()` which uses that guess to parse the column:
2015-09-23 21:58:16 +08:00
2016-07-07 02:59:50 +08:00
```{r}
2016-07-10 22:19:56 +08:00
guess_parser("2010-10-01")
guess_parser("15:01")
guess_parser(c("TRUE", "FALSE", "FALSE", "TRUE"))
guess_parser(c("1", "5", "9"))
2016-07-07 02:59:50 +08:00
```
2015-09-23 21:58:16 +08:00
2016-07-12 04:38:39 +08:00
The basic rules try each of the following rules in turn, working from strictest to most flexible:
2016-07-08 00:17:11 +08:00
2016-07-12 04:38:39 +08:00
* logical: contains only "F", "T", "FALSE", or "TRUE".
* integer: contains only numeric characters (and `-`).
* double: contains only valid doubles (including numbers like `4.5e-5`).
* number: contains valid doubles with the grouping mark inside.
* time: matches the default time format.
* date: matches the default date format.
* date time: any ISO8601 date.
2015-09-23 21:58:16 +08:00
2016-07-12 04:38:39 +08:00
If none of these rules apply, then it will get read in as a character vector. (Note that the details will change a little from version to version as we tweak the guesses to provide the best balance between false positives and false negatives)
2016-07-07 02:59:50 +08:00
2016-07-10 22:19:56 +08:00
### Problems
2016-07-07 02:59:50 +08:00
2016-07-10 22:19:56 +08:00
These defaults don't always work for larger files. There are two basic problems:
2016-07-07 02:59:50 +08:00
2016-07-10 22:19:56 +08:00
1. The first thousand rows might be a special case, and readr guesses
2016-07-12 04:38:39 +08:00
a type that is not sufficiently general. For example, you might have
a column of doubles that only contains integers in the first 1000 rows.
2016-07-07 02:59:50 +08:00
2016-07-10 22:19:56 +08:00
1. The column might contain a lot of missing values. If the first 1000
rows contains on `NA`s, readr will guess that it's a character
2016-07-12 04:38:39 +08:00
vector, whereas you probably want to parse it as something more
2016-07-10 22:19:56 +08:00
specific.
2016-07-07 02:59:50 +08:00
2016-07-10 22:19:56 +08:00
readr contains a challenging csv that illustrates both of these problems:
2016-07-07 02:59:50 +08:00
2016-07-10 22:19:56 +08:00
```{r}
challenge <- read_csv(readr_example("challenge.csv"))
2016-07-07 02:59:50 +08:00
```
2016-07-12 04:38:39 +08:00
(Note the use of `readr_example()` which finds the path to one of the files included with the package)
There are two outputs: the column specification generated by looking at the first 1000 rows, and the first five parsing failures. It's always a good idea to explicitly pull out the `problems()` so you can explore them in more depth:
2016-07-07 02:59:50 +08:00
2016-07-10 22:19:56 +08:00
```{r}
problems(challenge)
```
2016-01-27 21:33:35 +08:00
2016-07-12 04:38:39 +08:00
A good strategy is to work column by column until there are no problems remaining. Here we can see that there are a lot of parsing problems with the `x` column - there are trailing characters after the integer value. That suggests we need to use a double parser instead.
2016-07-07 02:59:50 +08:00
2016-07-12 04:38:39 +08:00
To fix the call, start by copying and pasting the column specification into your original call:
2016-07-07 02:59:50 +08:00
2016-07-10 22:19:56 +08:00
```{r, eval = FALSE}
challenge <- read_csv(
readr_example("challenge.csv"),
col_types = cols(
x = col_integer(),
y = col_character()
)
)
```
2016-07-07 02:59:50 +08:00
2016-07-10 22:19:56 +08:00
Then you can tweak the type of the `x` column:
2016-07-07 02:59:50 +08:00
2016-07-10 22:19:56 +08:00
```{r}
challenge <- read_csv(
readr_example("challenge.csv"),
col_types = cols(
x = col_double(),
y = col_character()
)
)
```
2016-07-07 02:59:50 +08:00
2016-07-10 22:19:56 +08:00
That fixes the first problem, but if we look at the last few rows, you'll see that they're dates stored in a character vector:
```{r}
tail(challenge)
```
You can fix that by specifying that `y` is date column:
```{r}
challenge <- read_csv(
readr_example("challenge.csv"),
col_types = cols(
x = col_double(),
y = col_date()
)
)
tail(challenge)
2016-07-07 02:59:50 +08:00
```
2016-07-10 22:19:56 +08:00
Every `parse_xyz()` function has a corresponding `col_xyz()` function. You use `parse_xyz()` when the data is in a character vector in R already; you use `col_xyz()` when you want to tell readr how to load the data.
2016-07-12 04:38:39 +08:00
I highly recommend building up a complete column specification using the print-out provided by readr. This ensures that you have a consistent, reproducible, data import script. If you rely on the default guesses and your data changes, readr will continue to read it in. If you want to be really strict, use `stop_for_problems()`: that function throws an error and stops your script if there are any parsing problems.
### Other strategies
2016-07-10 22:19:56 +08:00
2016-07-12 04:38:39 +08:00
There are a few other general strategies to help you parse files:
2016-07-07 02:59:50 +08:00
2016-07-10 22:19:56 +08:00
* In this case we just got unlucky, and if we'd looked at just
a few more rows, we could have correctly parsed in one shot:
```{r}
challenge2 <- read_csv(readr_example("challenge.csv"), guess_max = 1001)
challenge2
```
2016-07-07 02:59:50 +08:00
2016-07-10 22:19:56 +08:00
* Sometimes it's easier to diagnose problems if you just read in all
the columns as character vectors:
```{r}
challenge2 <- read_csv(readr_example("challenge.csv"),
col_types = cols(.default = col_character())
)
```
This is particularly useful in conjunction with `type_convert()`,
which applies the parsing heuristics to the character columns in a data
2016-07-12 04:38:39 +08:00
frame.
2016-07-07 02:59:50 +08:00
2016-07-10 22:19:56 +08:00
```{r}
df <- tibble::tibble(
x = c("1", "2", "3"),
y = c("1.21", "2.32", "4.56")
)
df
# Note the column types
type_convert(df)
```
* If you're reading a very large file, you might want to set `n_max` to
2016-07-12 04:38:39 +08:00
a smallish numberl like 10,000 or 100,000. That will speed up iteration
while you eliminate common problems.
2016-07-07 02:59:50 +08:00
2016-07-10 22:19:56 +08:00
* If you're having major parsing problems, sometimes it's easier
to just read into a character vector of lines with `read_lines()`,
or even a character vector of length 1 with `read_file()`. Then you
can use the string parsing skills you'll learn later to parse
more exotic formats.
2016-01-27 21:33:35 +08:00
2016-07-10 22:19:56 +08:00
## Writing to a file
2015-09-23 21:58:16 +08:00
2016-07-10 22:19:56 +08:00
readr also comes with two useful functions for writing data back to disk: `write_csv()` and `write_tsv()`. They:
2016-07-07 02:59:50 +08:00
2016-07-10 22:19:56 +08:00
* Are faster than the base R equvalents.
2016-07-07 02:59:50 +08:00
2016-07-10 22:19:56 +08:00
* Never write rownames, and quote only when needed.
2016-07-07 02:59:50 +08:00
2016-07-12 04:38:39 +08:00
* Always encode strings in UTF-8. If you want to export a csv file to
2016-07-10 22:19:56 +08:00
Excel, use `write_excel_csv()` - this writes a special character
2016-07-12 04:38:39 +08:00
(a "byte order mark") at the start of the file which tells Excel that
you're using the UTF-8 encoding.
2016-07-10 22:19:56 +08:00
* Save dates and datetimes in ISO8601 format so they are easily
parsed elsewhere.
2016-07-07 02:59:50 +08:00
2016-07-10 22:19:56 +08:00
The most important arguments are `x` (the data frame to save), and `path` (the location to save it). You can also specify how missing values are written with `na`, and if you want to `append` to an existing file.
2016-07-07 02:59:50 +08:00
2016-07-10 22:19:56 +08:00
```{r, eval = FALSE}
write_csv(challenge, "challenge.csv")
```
2016-07-07 02:59:50 +08:00
2016-07-10 22:19:56 +08:00
Note that the type information is lost when you save to csv:
2016-07-07 02:59:50 +08:00
2016-07-10 22:19:56 +08:00
```{r, warning = FALSE}
challenge
write_csv(challenge, "challenge-2.csv")
read_csv("challenge-2.csv")
2016-07-07 02:59:50 +08:00
```
2015-09-23 21:58:16 +08:00
2016-07-12 04:38:39 +08:00
This makes csvs a little unreliable for caching interim results - you need to recreate the column specification every time you load in. There are two alternatives:
2015-09-23 21:58:16 +08:00
2016-07-12 04:38:39 +08:00
1. `write_rds()` and `read_rds()` are uniform wrappers around the base
functions `readRDS()` and `saveRDS()`. These store data in R's custom
binary format:
2016-07-10 22:19:56 +08:00
```{r}
write_rds(challenge, "challenge.rds")
read_rds("challenge.rds")
```
1. The feather package implements a fast binary file format that can
be shared across programming languages:
```{r, eval = FALSE}
2016-07-10 22:19:56 +08:00
library(feather)
write_feather(challenge, "challenge.feather")
read_feather("challenge.feather")
#> # A tibble: 2,000 x 2
#> x y
#> <dbl> <date>
#> 1 404 <NA>
#> 2 4172 <NA>
#> 3 3004 <NA>
#> 4 787 <NA>
#> 5 37 <NA>
#> 6 2332 <NA>
#> # ... with 1,994 more rows
2016-07-10 22:19:56 +08:00
```
2015-09-21 21:41:14 +08:00
2016-07-12 04:38:39 +08:00
feather tends to be faster than rds and is usable outside of R. `rds` supports list-columns (which you'll learn about in [[Many models]]), which feather currently does not.
2015-09-21 21:41:14 +08:00
2016-07-10 22:19:56 +08:00
```{r, include = FALSE}
file.remove("challenge-2.csv")
file.remove("challenge.rds")
```
2015-09-21 21:41:14 +08:00
2016-07-07 02:59:50 +08:00
## Other types of data
2016-04-02 01:33:51 +08:00
2016-07-12 04:38:39 +08:00
To get other types of data into R, we recommend starting with the tidyverse packages listed below. They're certainly not perfect, but they are a good place to start.
2016-04-02 01:33:51 +08:00
For rectanuglar data:
2016-04-02 01:33:51 +08:00
* haven reads SPSS, Stata, and SAS files.
2016-04-02 01:33:51 +08:00
2016-07-07 02:59:50 +08:00
* readxl reads excel files (both `.xls` and `.xlsx`).
2016-04-02 01:33:51 +08:00
* DBI, along with a database specific backend (e.g. RMySQL, RSQLite,
RPostgreSQL etc) allows you to run SQL queries against a database
and return a data frame.
For hierarchical data:
* jsonlite (by Jeroen Ooms) reads json
2016-04-02 01:33:51 +08:00
* xml2 reads XML.
2016-07-12 04:38:39 +08:00
For more exotic file types, try the [R data import/export manual](https://cran.r-project.org/doc/manuals/r-release/R-data.html) and the [rio](https://github.com/leeper/rio) package.