functions_common.r

## math

abs(-1)
round(rnorm(1),2)
sum(rnorm(5))
mean(rnorm(10))
sd(rnorm(10))

## serial paste:
paste0("year_", 1967:1974)

## min max index
v1 <- sample(100,10)
v1[which.min(v1)]
v1[which.max(v1)]


## seq and rep

seq(8,2,by=-2)                          # start, stop, by
rep(seq(8,2,by=-2),times=2)             # twice the above
rep(seq(8,2,by=-2),each=2)              # each twice
sort(rep(seq(8,2,by=-2),times=2))       # the above sorted
sort(rep(seq(8,2,by=-2),times=2),decreasing=T)       # reverse sort

## range
?range                                  # basically, the "min and max" function
range(10)                               # 10 10
range(1,10)                             # 1 10
range(c(1:10))                          # 1 10
range(sample(100,10))                   # e.g. 17 92

## is.*() to check if it is, as.*() represent (cast) as another

is.list(list())                         # T
as.character(42)                        # "42"
is.character(as.character(42))          # T

list(1:3)                               # first element is a vector
as.list(1:3)                            # each list item is a one length vector

unlist(list(1:3))                       # [1] 1 2 3
unlist(as.list(1:3))                    # same as above,
## if several types are present, they are converted:
unlist(list(F,c(1:5),"name"))           # [1] "FALSE" "1"     "2"     "3"     "4"     "5"     "name"

## append

v <- 1:5
w <- 6:9
v
w
c(v,w)
append(v,w)                             # same as previous in a fancier way

## revert
rev(v)                                  # [1] 5 4 3 2 1
## string revert also in common.r:
(rev(strsplit("alpha",split="")[[1]]),collapse="")

## which

x <- c( 1, 2, NA, 3, NA)
y <- c(NA, 3, NA, 3,  4)

which(is.na(x) & is.na(y))              # find where both are NA

## na.omit
na.omit(x)
class(na.omit(y))

## match

v1 <- c("a","b","c")
"a" %in% v1                             # T
"d" %in% v1                             # F

## regex functions

## grep and grepl
source("common.r")
cit10 <- us.cities(10)
cit10
grepl("e.",cit10,ignore.case=T)         # logical vector
cit10[grepl("e.",cit10,ignore.case=T)]  # filter for the finds
cit10[grepl("^e.",cit10,ignore.case=T)] # first
cit10[grepl(".e",cit10,ignore.case=T)]  # not first
cit10[grepl(".e$",cit10,ignore.case=T)] # last
cit10[grepl("\\s",cit10,ignore.case=T)] # having a space

## ?regex                                  # for help

## grep gives index numbers:
grep("e.",cit10,ignore.case=T)         # e is not last
cit10[grep("e.",cit10,ignore.case=T)]
cit10[grep("(.).*\\1",cit10,ignore.case=T,perl=T)] # repeated letter
cit10[grep("(.)\\1",cit10,ignore.case=T,perl=T)] # double letter


## sub and gsub
cit10
gsub("[AEIOUaeiou]","*",cit10)               # replacement of vowels
gsub("[AEIOUaeiou]",sample(c("a","e","i","o","u"),1),cit10) # replacement of vowels 2

## date and time
my.time <- as.POSIXct("2016-07-11 23:34:10")
my.time
class(my.time)                          # "POSIXct" "POSIXt"
my.time + 1                             # seconds are incremented
unclass(my.time)                        # seconds since start of epoch time,
                                        # as first element of a list

my.date <- as.Date("2016-07-11")
my.date + 1                             # days are incremented
class(my.date)                          # date
typeof(my.date)                         # doble
unclass(my.date)                        # days since start of epoch

format(my.date,"%d %B %Y")

date1 <- as.Date("31 August 2015",format="%d %B %Y")
date1

unclass(Sys.time())

diff(date.generator(5))                 # time diferences in days
diff(time.generator(5))

mean(diff(date.generator(5)))           # average time difference

?as.Date

## packages to check out:
## lubridate
## zoo
## xts

## Warnings and messages, error:

## 1. warnings:
mean.ci <- function(x, level = 0.95) {
  if (length(x) == 0) {
    warning("`x` was empty", call. = FALSE)
    return(c(-Inf,Inf))
    }
  se <- sd(x) / sqrt(length(x))
  alpha <- 1 - level
  mean(x) + se * qnorm(c(alpha / 2, 1 - alpha / 2))
}

v1 <- vector("numeric",0)

mean.ci(v1)

## replace NA
x = as.numeric(c(1,2,NA,4,5))

x[is.na(x)] <- 3

x                                       # 3 is there in the middle

## 2. messages
message("one ","two ",x)
## example:
replace.missings <- function(x, replacement) {
  is_miss <- is.na(x)
  x[is_miss] <- replacement
  
  # Rewrite to use message()
  message(sum(is_miss)," missings replaced by ", replacement, "\n")
  x
}

v1 <- sample(c(1:5,NA),10,replace=T)
v1
replace.missings(v1,0)


## 3. Error (best inside a function or module)

stopifnot(is.character(x))
## which is short for:
if (!is.character(x))
    stop("`x` should be a character vector", call. = F)

?quantile