Basic usage of R

# Basic usage of R - 3

## Introductory Computer Programming

### Deepayan Sarkar

---

# Functions

---

* Most useful things in R happen by calling functions

* Functions have one or more arguments

* All arguments have names (available as variables inside the function)
--

* Arguments may be compulsory or optional

* Optional arguments usually have "default" values 
--

* Arguments may or may not be named when the function is called
	
	* Unnamed arguments are matched by position

* Optional arguments are usually named
	
* Functions normally also have a useful "return" value

---

* For example, linear models are fit using the function `lm()`

```r
str(lm)
```

```
function (formula, data, subset, weights, na.action, method = "qr", model = TRUE, x = FALSE, 
    y = FALSE, qr = TRUE, singular.ok = TRUE, contrasts = NULL, offset, ...)  
```

* These calls are equivalent

```r
fm1 <- lm(mpg ~ wt, mtcars, (am == 0))
fm2 <- lm(mpg ~ wt, data = mtcars, method = "qr", subset = (am == 0))
coef(fm1)
```

```
(Intercept)          wt 
  31.416055   -3.785908 
```

```r
coef(fm2)
```

```
(Intercept)          wt 
  31.416055   -3.785908 
```

---

* Rule: named arguments are matched by name, remaining by position

* Convention:

- Named arguments are usually not followed by unnamed arguments

- First few "standard" arguments are usually unnamed (matched by position)
	
* The following call is equivalent to previous two, but not recommended

```r
fm3 <- lm(mpg ~ wt, data = mtcars, (am == 0))
coef(fm3)
```

```
(Intercept)          wt 
  31.416055   -3.785908 
```

---

* The return value of `lm()` is a list

```r
str(fm2)
```

```
List of 12
 $ coefficients : Named num [1:2] 31.42 -3.79
  ..- attr(*, "names")= chr [1:2] "(Intercept)" "wt"
 $ residuals    : Named num [1:19] 2.156 0.307 -0.217 -3.6 5.061 ...
  ..- attr(*, "names")= chr [1:19] "Hornet 4 Drive" "Hornet Sportabout" "Valiant" "Duster 360" ...
 $ effects      : Named num [1:19] -74.744 -12.487 -0.614 -4 4.672 ...
  ..- attr(*, "names")= chr [1:19] "(Intercept)" "wt" "" "" ...
 $ rank         : int 2
 $ fitted.values: Named num [1:19] 19.2 18.4 18.3 17.9 19.3 ...
  ..- attr(*, "names")= chr [1:19] "Hornet 4 Drive" "Hornet Sportabout" "Valiant" "Duster 360" ...
 $ assign       : int [1:2] 0 1
 $ qr           :List of 5
  ..$ qr   : num [1:19, 1:2] -4.359 0.229 0.229 0.229 0.229 ...
  .. ..- attr(*, "dimnames")=List of 2
  .. .. ..$ : chr [1:19] "Hornet 4 Drive" "Hornet Sportabout" "Valiant" "Duster 360" ...
  .. .. ..$ : chr [1:2] "(Intercept)" "wt"
  .. ..- attr(*, "assign")= int [1:2] 0 1
  ..$ qraux: num [1:2] 1.23 1.07
  ..$ pivot: int [1:2] 1 2
  ..$ tol  : num 1e-07
  ..$ rank : int 2
  ..- attr(*, "class")= chr "qr"
 $ df.residual  : int 17
 $ xlevels      : Named list()
 $ call         : language lm(formula = mpg ~ wt, data = mtcars, subset = (am == 0), method = "qr")
 $ terms        :Classes 'terms', 'formula'  language mpg ~ wt
  .. ..- attr(*, "variables")= language list(mpg, wt)
  .. ..- attr(*, "factors")= int [1:2, 1] 0 1
  .. .. ..- attr(*, "dimnames")=List of 2
  .. .. .. ..$ : chr [1:2] "mpg" "wt"
  .. .. .. ..$ : chr "wt"
  .. ..- attr(*, "term.labels")= chr "wt"
  .. ..- attr(*, "order")= int 1
  .. ..- attr(*, "intercept")= int 1
  .. ..- attr(*, "response")= int 1
  .. ..- attr(*, ".Environment")=<environment: R_GlobalEnv> 
  .. ..- attr(*, "predvars")= language list(mpg, wt)
  .. ..- attr(*, "dataClasses")= Named chr [1:2] "numeric" "numeric"
  .. .. ..- attr(*, "names")= chr [1:2] "mpg" "wt"
 $ model        :'data.frame':	19 obs. of  2 variables:
  ..$ mpg: num [1:19] 21.4 18.7 18.1 14.3 24.4 22.8 19.2 17.8 16.4 17.3 ...
  ..$ wt : num [1:19] 3.21 3.44 3.46 3.57 3.19 ...
  ..- attr(*, "terms")=Classes 'terms', 'formula'  language mpg ~ wt
  .. .. ..- attr(*, "variables")= language list(mpg, wt)
  .. .. ..- attr(*, "factors")= int [1:2, 1] 0 1
  .. .. .. ..- attr(*, "dimnames")=List of 2
  .. .. .. .. ..$ : chr [1:2] "mpg" "wt"
  .. .. .. .. ..$ : chr "wt"
  .. .. ..- attr(*, "term.labels")= chr "wt"
  .. .. ..- attr(*, "order")= int 1
  .. .. ..- attr(*, "intercept")= int 1
  .. .. ..- attr(*, "response")= int 1
  .. .. ..- attr(*, ".Environment")=<environment: R_GlobalEnv> 
  .. .. ..- attr(*, "predvars")= language list(mpg, wt)
  .. .. ..- attr(*, "dataClasses")= Named chr [1:2] "numeric" "numeric"
  .. .. .. ..- attr(*, "names")= chr [1:2] "mpg" "wt"
 - attr(*, "class")= chr "lm"
```

]

---

* Individual elements can be extracted using list indexing

```r
fm2$coefficients
```

```
(Intercept)          wt 
  31.416055   -3.785908 
```

```r
fm2$residuals
```

```
     Hornet 4 Drive   Hornet Sportabout             Valiant          Duster 360           Merc 240D 
        2.155637322         0.307466517        -0.216815332        -3.600365503         5.060989634 
           Merc 230            Merc 280           Merc 280C          Merc 450SE          Merc 450SL 
        3.309553333         0.807466517        -0.592533483         0.392588263         0.005379702 
        Merc 450SLC  Cadillac Fleetwood Lincoln Continental   Chrysler Imperial       Toyota Corona 
       -1.905324922        -1.140040848        -0.481292938         3.519620367        -0.583793327 
   Dodge Challenger         AMC Javelin          Camaro Z28    Pontiac Firebird 
       -2.589660880        -3.211463020        -3.578170470         2.340759068 
```

---

# Attributes

---

* The names (derived from row names of the data) can be used as index

```r
fm2$residuals["Chrysler Imperial"]
```

```
Chrysler Imperial 
          3.51962 
```

* The names associated with a vector can be obtained using `names()`

```r
names(fm2$residuals)
```

```
 [1] "Hornet 4 Drive"      "Hornet Sportabout"   "Valiant"             "Duster 360"         
 [5] "Merc 240D"           "Merc 230"            "Merc 280"            "Merc 280C"          
 [9] "Merc 450SE"          "Merc 450SL"          "Merc 450SLC"         "Cadillac Fleetwood" 
[13] "Lincoln Continental" "Chrysler Imperial"   "Toyota Corona"       "Dodge Challenger"   
[17] "AMC Javelin"         "Camaro Z28"          "Pontiac Firebird"   
```

---

* These names are actually stored as an _attribute_ called "names"

```r
attr(fm2$residuals, "names")
```

* Attributes are arbitrary R objects that can be attached to any other object

* Typically used for programming convenience, normally not seen by users

* However, some attributes are "special"

---

* The "names" attribute can be extracted using the function `names()`

* `dimnames()` similarly gives row / column names for matrices and arrays

```r
dimnames(VADeaths)
```

```
[[1]]
[1] "50-54" "55-59" "60-64" "65-69" "70-74"

[[2]]
[1] "Rural Male"   "Rural Female" "Urban Male"   "Urban Female"
```

```r
attr(VADeaths, "dimnames")
```

```
[[1]]
[1] "50-54" "55-59" "60-64" "65-69" "70-74"

[[2]]
[1] "Rural Male"   "Rural Female" "Urban Male"   "Urban Female"
```

---

* For example, column names can be obtained as

```r
dimnames(VADeaths)[[2]]
```

```
[1] "Rural Male"   "Rural Female" "Urban Male"   "Urban Female"
```

* There are convenient shortcuts called `rownames()` and `colnames()`

```r
colnames(VADeaths)
```

```
[1] "Rural Male"   "Rural Female" "Urban Male"   "Urban Female"
```

---

* In fact, we can easily verify that this is what `colnames()` is doing by printing it

```r
colnames
```

```
function (x, do.NULL = TRUE, prefix = "col") 
{
    if (is.data.frame(x) && do.NULL) 
        return(names(x))
    dn <- dimnames(x)
    if (!is.null(dn[[2L]])) 
        dn[[2L]]
    else {
        nc <- NCOL(x)
        if (do.NULL) 
            NULL
        else if (nc > 0L) 
            paste0(prefix, seq_len(nc))
        else character()
    }
}
<bytecode: 0x7f92327ceb68>
<environment: namespace:base>
```

* All R functions can be easily inspected in this way

---

* Another very important attribute is "class"

* For example, the return value of `lm()` has class "lm"

```r
attr(fm2, "class")
```

```
[1] "lm"
```

* The class of an object can also be obtained using the function `class()`

```r
class(fm2)
```

```
[1] "lm"
```

---

* The class of an object can (usually) be "removed" by setting it to
  `NULL` 
  
* This is not something you should actually do!

```r
class(fm2) <- NULL
```

* Such objects will no longer have a "class" attribute

```r
attr(fm2, "class")
```

```
NULL
```

* However, it will still have a class (implicitly)

```r
class(fm2)
```

```
[1] "list"
```
--

* We will discuss the class attribute in more details next day

---

# Questions?