Choose names, variables, cases - base R

TODO

• link to strings for grep(), dfSplitMerge, dfReshape

Simulate data

set.seed(123)
N      <- 12
sex    <- factor(sample(c("f", "m"), N, replace=TRUE), levels=c("f", "m"))
group  <- factor(sample(rep(c("CG", "WL", "T"), 4), N, replace=FALSE), levels=c("CG", "WL", "T"))
age    <- sample(18:35, N, replace=TRUE)
IQ     <- round(rnorm(N, mean=100, sd=15))
rating <- round(runif(N, min=0, max=6))
(myDf1 <- data.frame(id=1:N, sex, group, age, IQ, rating))

   id sex group age  IQ rating
1   1   f     T  25  95      5
2   2   f     T  24  84      5
3   3   f    CG  27  99      3
4   4   m    WL  26 116      5
5   5   f     T  21  98      4
6   6   m    WL  31  83      4
7   7   m    CG  34  88      0
8   8   m    CG  28 110      3
9   9   f     T  24  95      1
10 10   f    WL  29  80      2
11 11   m    CG  32  91      4
12 12   m    WL  27  98      2

Rename variables

names(myDf1)

[1] "id"     "sex"    "group"  "age"    "IQ"     "rating"

names(myDf1)[3] <- "fac"
names(myDf1)

[1] "id"     "sex"    "fac"    "age"    "IQ"     "rating"

names(myDf1)[names(myDf1) == "fac"] <- "group"
names(myDf1)

[1] "id"     "sex"    "group"  "age"    "IQ"     "rating"

Select variables

Select by full name

subset(myDf1, select=c(group, IQ))

   group  IQ
1      T  95
2      T  84
3     CG  99
4     WL 116
5      T  98
6     WL  83
7     CG  88
8     CG 110
9      T  95
10    WL  80
11    CG  91
12    WL  98

Select by name pattern

(colIdx <- grep("^i.+", names(myDf1), ignore.case=TRUE))

[1] 1 5

subset(myDf1, select=colIdx)

   id  IQ
1   1  95
2   2  84
3   3  99
4   4 116
5   5  98
6   6  83
7   7  88
8   8 110
9   9  95
10 10  80
11 11  91
12 12  98

Select by data type

Only numeric variables

(colIdx <- sapply(myDf1, is.numeric))

    id    sex  group    age     IQ rating 
  TRUE  FALSE  FALSE   TRUE   TRUE   TRUE 

subset(myDf1, select=colIdx)

   id age  IQ rating
1   1  25  95      5
2   2  24  84      5
3   3  27  99      3
4   4  26 116      5
5   5  21  98      4
6   6  31  83      4
7   7  34  88      0
8   8  28 110      3
9   9  24  95      1
10 10  29  80      2
11 11  32  91      4
12 12  27  98      2

More elegant

myDfNum <- Filter(is.numeric, myDf1)
head(myDfNum)

  id age  IQ rating
1  1  25  95      5
2  2  24  84      5
3  3  27  99      3
4  4  26 116      5
5  5  21  98      4
6  6  31  83      4

De-select variables

subset(myDf1, select=c(-sex, -IQ))

   id group age rating
1   1     T  25      5
2   2     T  24      5
3   3    CG  27      3
4   4    WL  26      5
5   5     T  21      4
6   6    WL  31      4
7   7    CG  34      0
8   8    CG  28      3
9   9     T  24      1
10 10    WL  29      2
11 11    CG  32      4
12 12    WL  27      2

Remove variables - 1 variable

dfTemp <- myDf1
dfTemp$group <- NULL
head(dfTemp, n=3)

  id sex age IQ rating
1  1   f  25 95      5
2  2   f  24 84      5
3  3   f  27 99      3

Remove variables - multiple variables

dfTemp[c("sex", "IQ")] <- list(NULL)
head(dfTemp, n=3)

  id age rating
1  1  25      5
2  2  24      5
3  3  27      3

Select cases

Select by logical conditions

Simple conditions

subset(myDf1, sex == "f")

   id sex group age IQ rating
1   1   f     T  25 95      5
2   2   f     T  24 84      5
3   3   f    CG  27 99      3
5   5   f     T  21 98      4
9   9   f     T  24 95      1
10 10   f    WL  29 80      2

subset(myDf1, id == rating)

  id sex group age IQ rating
3  3   f    CG  27 99      3

Combined conditions - logical AND

subset(myDf1, (sex == "m") & (rating > 2))

   id sex group age  IQ rating
4   4   m    WL  26 116      5
6   6   m    WL  31  83      4
8   8   m    CG  28 110      3
11 11   m    CG  32  91      4

Combined conditions - logical OR

subset(myDf1, (IQ < 90) | (IQ > 110))

   id sex group age  IQ rating
2   2   f     T  24  84      5
4   4   m    WL  26 116      5
6   6   m    WL  31  83      4
7   7   m    CG  34  88      0
10 10   f    WL  29  80      2

subset(myDf1, group %in% c("CG", "WL"))

   id sex group age  IQ rating
3   3   f    CG  27  99      3
4   4   m    WL  26 116      5
6   6   m    WL  31  83      4
7   7   m    CG  34  88      0
8   8   m    CG  28 110      3
10 10   f    WL  29  80      2
11 11   m    CG  32  91      4
12 12   m    WL  27  98      2

Identify and drop duplicates

myDfDouble <- rbind(myDf1, myDf1[sample(seq_len(nrow(myDf1)), 4), ])
duplicated(myDfDouble) | duplicated(myDfDouble, fromLast=TRUE)

 [1] FALSE FALSE FALSE  TRUE FALSE  TRUE FALSE  TRUE FALSE FALSE  TRUE FALSE
[13]  TRUE  TRUE  TRUE  TRUE

myDfUnique <- unique(myDfDouble)
any(duplicated(myDfUnique))

[1] FALSE

Identify, count, and drop cases with missings

myDfNA           <- myDf1
myDfNA$IQ[2]     <- NA
myDfNA$rating[3] <- NA

Count missings

is.na(myDfNA)[1:3, ]

        id   sex group   age    IQ rating
[1,] FALSE FALSE FALSE FALSE FALSE  FALSE
[2,] FALSE FALSE FALSE FALSE  TRUE  FALSE
[3,] FALSE FALSE FALSE FALSE FALSE   TRUE

apply(myDfNA, 2, anyNA)

    id    sex  group    age     IQ rating 
 FALSE  FALSE  FALSE  FALSE   TRUE   TRUE 

Count missings / present observations by variable

apply(myDfNA, 2, function(x) { sum(is.na(x)) })

    id    sex  group    age     IQ rating 
     0      0      0      0      1      1 

apply(myDfNA, 2, function(x) { length(na.omit(x)) })

    id    sex  group    age     IQ rating 
    12     12     12     12     11     11 

Drop cases with missings

complete.cases(myDfNA)

 [1]  TRUE FALSE FALSE  TRUE  TRUE  TRUE  TRUE  TRUE  TRUE  TRUE  TRUE  TRUE

head(na.omit(myDfNA), n=4)

  id sex group age  IQ rating
1  1   f     T  25  95      5
4  4   m    WL  26 116      5
5  5   f     T  21  98      4
6  6   m    WL  31  83      4

table(complete.cases(myDfNA))

FALSE  TRUE 
    2    10 

Show only cases with missings

subset(myDfNA, !complete.cases(myDfNA))

  id sex group age IQ rating
2  2   f     T  24 NA      5
3  3   f    CG  27 99     NA

Change data frame variables

Add new variables or overwrite existing variables

married <- sample(c(TRUE, FALSE), nrow(myDf1), replace=TRUE)
myDf2   <- myDf1
myDf2$married1    <- married
myDf2["married2"] <- married
head(myDf2)

  id sex group age  IQ rating married1 married2
1  1   f     T  25  95      5    FALSE    FALSE
2  2   f     T  24  84      5     TRUE     TRUE
3  3   f    CG  27  99      3     TRUE     TRUE
4  4   m    WL  26 116      5    FALSE    FALSE
5  5   f     T  21  98      4     TRUE     TRUE
6  6   m    WL  31  83      4    FALSE    FALSE

myDf3 <- cbind(myDf1, married)
head(myDf3, n=3)

  id sex group age IQ rating married
1  1   f     T  25 95      5   FALSE
2  2   f     T  24 84      5    TRUE
3  3   f    CG  27 99      3    TRUE

myDf4 <- transform(myDf3,
                   rSq=rating^2,
                   IQgrp=cut(IQ, breaks=c(0, 100, Inf)))
head(myDf4, n=3)

  id sex group age IQ rating married rSq   IQgrp
1  1   f     T  25 95      5   FALSE  25 (0,100]
2  2   f     T  24 84      5    TRUE  25 (0,100]
3  3   f    CG  27 99      3    TRUE   9 (0,100]

Change order of rows

(idx1 <- order(myDf1$rating))

 [1]  7  9 10 12  3  8  5  6 11  1  2  4

myDf1[idx1, ]

   id sex group age  IQ rating
7   7   m    CG  34  88      0
9   9   f     T  24  95      1
10 10   f    WL  29  80      2
12 12   m    WL  27  98      2
3   3   f    CG  27  99      3
8   8   m    CG  28 110      3
5   5   f     T  21  98      4
6   6   m    WL  31  83      4
11 11   m    CG  32  91      4
1   1   f     T  25  95      5
2   2   f     T  24  84      5
4   4   m    WL  26 116      5

(idx2 <- order(myDf1$group, myDf1$IQ))

 [1]  7 11  3  8 10  6 12  4  2  1  9  5

myDf1[idx2, ]

   id sex group age  IQ rating
7   7   m    CG  34  88      0
11 11   m    CG  32  91      4
3   3   f    CG  27  99      3
8   8   m    CG  28 110      3
10 10   f    WL  29  80      2
6   6   m    WL  31  83      4
12 12   m    WL  27  98      2
4   4   m    WL  26 116      5
2   2   f     T  24  84      5
1   1   f     T  25  95      5
9   9   f     T  24  95      1
5   5   f     T  21  98      4

Useful packages

In order to work with data frames like with a database, use sqldf. You can then use standard SQL commands to select data. data.table provides an alternative (compatible) class of data frames with better performance for large amounts of data.

Get the article source from GitHub

R markdown - markdown - R code - all posts