---
title: Length and names
author: "Eric C. Anderson"
output:
  html_document:
    toc: yes
  bookdown::html_chapter:
    toc: no
layout: default_with_disqus
---


```{r setup, echo=FALSE, include=FALSE}
# PLEASE DO NOT EDIT THIS CODE BLOCK
library(knitr)
library(rrhw)
# tell knitr where to find the inserted file in case
# jekyll is building this in the top directory of the repo
opts_knit$set(child.path = paste(prj_dir_containing("rep-res-course.Rproj"), "extras/knitr_children/", sep=""))

init_homework("Length and names lecture")
rr_github_name <- NA
rr_pull_request_time <- NA
rr_question_chunk_name <- "NotSet"
rr_branch_name <- "ex-test"
rr_hw_file_name <- "exercises/trial_homework.rmd"
```


# Object (vector) _attributes_.  _length_, _names_. {#attr-length-names} 

## Object Attributes {#object-attributes}

* Before we can talk about indexing with _names_ we have to talk about the fact that any object can possess _attributes_.
* Every R object has two _intrinsic_ attributes: `mode` and `length`
* For atomic vectors, possible modes are: (fill in the blanks)
* The functions `mode(x)` and `length(x)` return these attributes for `x`

### The _mode()_ function
```{r}
set.seed(6)
x <- sample(x = letters, size = 10)

# a vector of letters:
x
# what mode is it?
mode(x)

# a vector of complex numbers
y <- 1:10 + 0+3i
y

# what mode is it?
mode(y)
```

### The _length()_ function
* You are more likely to use `length()` than the `mode()` function.  In fact, you
will use it all the time.
* As you might guess, it returns the _length_ of an object (like a vector) as an integer.
    ```{r}
    x <- seq(1, 5, by=.67)
    x

    # how long is that?
    length(x)

    # how can we pick out just the last element?
    x[length(x)]

    # how can we return a vector of everything but the last element?
    x[-length(x)]
    ```

### The replacement form of the _length()_ function

* Check out this little bit of syntactic sugar: To change the length of an
object called `obj`, say, you can do like this:
```{r, eval=FALSE}
length(obj) <- 16
```
* It will chop off the end if the new value is smaller than the old value
* It will pad the end with NAs if the the new value is larger than the old value
    ```{r}
    x <- seq(6.5, 8.8, length.out = 15)
    x

    # make it longer
    length(x) <- 25    
    x

    # now, chop it off if you want:
    length(x) <- 7
    ``` 

## A few problems for thought

```{r middle-extracto, rr.question=TRUE}
# here is a vector:
y <- seq(pi, 15*pi, by=pi)

# give me all the elements from the 4th up to the 3rd from the last:

```


```{r, gimme-n-of-something, rr.question=TRUE}
set.seed(10)
# here are a certain number of letters:
alpha <- sample(letters, size = as.integer(runif(1, min=4, max=20)))

# simulate a random normal deviate for each one (using rnorm)

```





## The _names_ Attribute of a Vector {#names-attribute}

* R gives you the option of having a name for every element of a vector
* You can set the _names_ attribute of a vector with the _replacement form_ 
of the `names()` function.
* You can query the _names_ attribute with the `names()` function (it returns a character vector).
* You can index a vector that has a _names_ attribute with names!

### Setting the _names_ of a vector
```{r}
# here's a vector
x <- c(5,4,7,8)

# here we set its names to whatever we want
names(x) <- c("first", "second", "third", "boing")

# when we (auto)print the vector, the names are included above the value:
x
```

### Reading names in vector output

* This can take some getting used to. For the first 10 years I worked with R I always got confused
about what was a name and what was a value in the output.  
* And I thought it was friggin' ugly to have all those names on there sometime
* And it "can" strain the eyeballs if the names are long. Consider this ridiculous example:
    ```{r}
    # values = the first 17 values of the alphabet
    ab <- letters[1:17]
    ab  # not so hard to look at

    # the sha1 hashes for each of the 17 letters
    library(digest)
    hashes <- unname(sapply(ab, digest, algo = "sha1"))
    hashes  # sort of ugly to look at

    # name each element in ab with its hash
    names(ab) <- hashes
    ab  # print it out now...Really hard to look at

    # but note that we can index with the names (as strings)
    ab["221799200137b7d72dfc4a618465bec71333a58b"]

    # in the above, output, what is the value and what is the name?
    ```
    
### Get rid of those damn names so I can read the thing

Sometimes you just want to get rid of the names to read stuff, or you
might have another legitimate reason to do so.  A handy way to do this is
with the `unname()` function
```{r}
ab  # whoa horribly ugly named output

unname(ab)  # returns its argument, but with the names attribute stripped off
```

```{r}
x[c("boing", "second")] # note names are retained in result 

x[c("third", "boing", "oops", "first", "first")] # note NA in result
```


### Return the names as a vector
This is super important:

The names attribute of a vector is just a _vector_ itself of mode character.  
```{r}
x <- c(5,4,7,8)  # here is a vector

# here we give it a names attribute, thus providing a name
# for every element in it
names(x) <- c("first", "second", "third", "boing")
x # see it printed with the names

# now, we can do this to get the names back as a vector
names(x)

# super important!  We can even index it directly.
# for example: get all the names for which the value
# of x is greater than 6:
names(x)[x>6]
```

### You can index with names!
Though it might not seem, at first, to be super useful, this is
incredibly useful.  You can do something like this
`x[c("boing", "second")]` to extract elements from a named vector.

```{r}
# same setup as before:
x <- c(5,4,7,8) 
names(x) <- c("first", "second", "third", "boing")

# behold the power!
x[c("boing", "first")]
```



### Assignment Form names Indexing 

When you use the _assignment form of the indexing operator_, and you include a name that doesn't exist, it expands the vector beyond its current length
```{r}
x <- c(5,4,7,8)

names(x) <- c("first", "second", "third", "boing")
x

x[c("first", "third", "oofdah", "squawk")] <- c(-1,-2,-3,-4)
x
```

### Using names indexing as an associative array
The power of names indexing really comes through when you want to return a value for
every unique name or identifier. This sort of construction occurs all the time.

Here is a contrived example:  

* You are studying fish behavior and you have 16 fish 
that you have labelled `A` through ``r LETTERS[16]``, and you
have tagged them so you can tell who they are when you see them.
    ```{r}
    IDs <- LETTERS[1:16]
    ```
* You have recorded the fork length of each fish to the nearest mm
    ```{r}
    set.seed(16)
    fklen <- floor(rnorm(length(IDs), mean = 150, sd = 15))
    ```
* You also have watched them all day and you have recorded the order in which they have been
jumping all day long.  There have been 487 fish jumps recorded today and you have recorded
the sequence:
    ```{r}
    sequence <- sample(IDs, 487, replace = TRUE)
    ```

Now, what you are really interested in is whether big fish are more likely to jump after big fish than small fish, so you really need a vector which gives the sequence of the _fork lengths_ of the fish that were jumping.  Can you see how to do that using names?  Go for it.

### Other non-intrinsic Attributes (can be skipped for now...)

* Any object in R can have a number of different _attributes_ these are _not_ the data contained in the object, but they may affect how the object is treated.  
* Attributes are fairly central to the operation of R.
* Relevant functions:
    ```{r, eval=FALSE}
    attributes(x)  # list all non-intrinsic attributes of x
    
    attributes(x) <- value # set all attributes of x  (seldom used)
    attr(x, "boing") # return value of x's "boing" attribute
    attr(x, "boing") <- value # set x's "boing" attribute to value
    ```
Common attributes accessed via various convenience functions