table2df: Convert 'table' Objects to 'data.frame's

In mrdwab/mrdwabmisc: Miscellaneous R functions, mostly for data processing

Description

The table2df function takes an object of "class" table, ftable, and xtabs and converts them to data.frames, while retaining as many of the name details as possible.

Usage

table2df(mytable, as.multitable = FALSE, direction = "wide")

Arguments

mytable	The table object you want to convert into a data.frame. This can be an object in your workspace, or you can make the call to table, ftable, or xtabs as the mytable argument to this function.
as.multitable	Logical; defaults to FALSE. Some methods, for instance xtabs and table, will create an array of tables as the output when more than two variables are being tabulated. If as.multitable is TRUE, the function will return a list of data.frames. If as.multitable is FALSE, the function will convert the object to an ftable object before performing the transformation.
direction	Can be either "long" or "wide". If "long", the frequencies will all be tabulated into a single column. This is the same behavior you will generally get if you used as.data.frame on a table object. If "wide", the tabular format is retained.

Author(s)

Ananda Mahto

References

The expand.grid method for remaking the columns from an ftable was described by Kohske at http://stackoverflow.com/a/6463137/1270695.

See Also

table, ftable, xtabs

Examples

# Make up some data:
set.seed(1)
handedness <- data.frame(
gender = sample(c("Female", "Male", "Unknown"), 200, replace = TRUE),
handedness = sample(c("Right", "Left", "Ambidextrous"),
                 200, replace = TRUE, prob = c(.7, .2, .1)),
fav.col = sample(c("Red", "Orange", "Yellow", "Green", "Blue",
              "Indigo", "Violet", "Black", "White"),
              200, replace = TRUE),
fav.shape = sample(c("Triangle", "Circle", "Square", "Pentagon", "Hexagon",
                "Oval", "Octagon", "Rhombus", "Trapezoid"),
                200, replace = TRUE),
computer = sample(c("Win", "Mac", "Lin"), 200, replace = TRUE,
               prob = c(.5, .25, .25)))
# Preview the data
list(head(handedness), tail(handedness))

# A very basic table
HT1 <- with(handedness, table(gender, handedness))
HT1
table2df(HT1)
#'table2df(HT1, direction = "long")

# Another basic table
HT2 <- with(handedness, table(fav.col, computer))
HT2
table2df(HT2)

# This will create multiple tables, one for each possible computer value
HT3 <- with(handedness, table(gender, fav.col, computer))
HT3

# Default settings
table2df(HT3)

# As a list of data.frames
table2df(HT3, as.multitable = TRUE)

# As above, but with the output in long format
#   Only showing the first three lines of each data.frame
lapply(table2df(HT3, as.multitable = TRUE, direction = "long"), head, 3)

# Applied to an ftable
HT4 <- ftable(handedness,
         col.vars="fav.col",
         row.vars=c("gender", "computer"))
HT4
table2df(HT4)

# Applied to a single-row table
table2df(xtabs(breaks ~ wool, warpbreaks))

## ======================================= ##
## ========== OTHER EXAMPLES ============= ##

## Not run: 
table2df(xtabs(cbind(ncases, ncontrols) ~ ., data = esoph))
table2df(xtabs(cbind(ncases, ncontrols) ~ ., data = esoph),
    direction = "long")
table2df(xtabs(cbind(ncases, ncontrols) ~ ., data = esoph),
    as.multitable = TRUE, direction = "long")
table2df(xtabs(cbind(ncases, ncontrols) ~ ., data = esoph),
    as.multitable = TRUE, direction = "wide")

## End(Not run)

mrdwab/mrdwabmisc documentation built on May 23, 2019, 7:15 a.m.