Paste data.frame rows into an ordered factor

Usage

pasteByRowOrdered(
  x,
  sep = "_",
  na.rm = TRUE,
  condenseBlanks = TRUE,
  includeNames = FALSE,
  keepOrder = FALSE,
  byCols = seq_len(ncol(x)),
  na.last = TRUE,
  ...
)

Arguments

x: data.frame
sep: character separator to use between columns
na.rm: logical whether to remove NA values, or include them as "NA"
condenseBlanks: logical whether to condense blank or empty values without including an extra delimiter between columns.
includeNames: logical whether to include the colname delimited prior to the value, using sepName as the delimiter.
keepOrder: logical indicating whether non-factor columns should order factor levels based upon the existing order of unique items. This option is intended for data.frame whose columns are already sorted in proper order, but where columns are not factor with appropriate factor levels. Note that even when keepOrder=TRUE all existing factor columns will honor the order of factor levels already present in those columns.
byCols: integer or character passed to mixedSortDF(). This argument defines the order of columns sorted by mixedSortDF(), and does not affect the order of columns pasted. Columns are always pasted in the same order they appear in x. This argument byCols was previously passed via ... but is added here to make this connection more direct.
na.last: logical passed to base::factor() to determine whether NA values are first or last in factor level order.
...: additional arguments are passed to jamba::pasteByRow(), and to jamba::mixedSortDF().

Value

factor vector whose levels are defined by existing factor levels, then by sorted values.

Details

This function is an extension to jamba::pasteByRow() which pastes rows from a data.frame into a character vector. This function defines factor levels by running jamba::mixedSortDF(unique(x)) and calling jamba::pasteByRow() on the result. Therefore the original order of the input x is maintained while the factor levels are based upon the appropriate column-based sort.

Note that the ... additional arguments are passed to jamba::mixedSortDF() to customize the column-based sort order, used to define factor levels. A good way to test the order of factors is to run jamba::mixedSortDF(unique(x)) with appropriate arguments, and confirm the rows are ordered as expected.

Note also that jamba::mixedSortDF() uses jamba::mixedSort() which itself performs alphanumeric sort in order to keep values in proper numeric order where possible.

Examples

f <- LETTERS;
df <- data.frame(A=f[rep(1:3, each=2)],
   B=c(NA, f[3]),
   C=c(NA, NA, f[2]))
df
#>   A    B    C
#> 1 A <NA> <NA>
#> 2 A    C <NA>
#> 3 B <NA>    B
#> 4 B    C <NA>
#> 5 C <NA> <NA>
#> 6 C    C    B

# note that output is consistent with mixedSortDF()
jamba::mixedSortDF(df)
#>   A    B    C
#> 2 A    C <NA>
#> 1 A <NA> <NA>
#> 4 B    C <NA>
#> 3 B <NA>    B
#> 6 C    C    B
#> 5 C <NA> <NA>
jamba::pasteByRowOrdered(df)
#>     1     2     3     4     5     6 
#>     A   A_C   B_B   B_C     C C_C_B 
#> Levels: A_C A B_C B_B C_C_B C

jamba::mixedSortDF(df, na.last=FALSE)
#>   A    B    C
#> 1 A <NA> <NA>
#> 2 A    C <NA>
#> 3 B <NA>    B
#> 4 B    C <NA>
#> 5 C <NA> <NA>
#> 6 C    C    B
jamba::pasteByRowOrdered(df, na.last=FALSE)
#>     1     2     3     4     5     6 
#>     A   A_C   B_B   B_C     C C_C_B 
#> Levels: A A_C B_B B_C C C_C_B

jamba::mixedSortDF(df, byCols=c(3, 2, 1))
#>   A    B    C
#> 6 C    C    B
#> 3 B <NA>    B
#> 2 A    C <NA>
#> 4 B    C <NA>
#> 1 A <NA> <NA>
#> 5 C <NA> <NA>
jamba::pasteByRowOrdered(df, byCols=c(3, 2, 1))
#>     1     2     3     4     5     6 
#>     A   A_C   B_B   B_C     C C_C_B 
#> Levels: C_C_B B_B A_C B_C A C

df1 <- data.frame(group=rep(c("Control", "ABC1"), each=6),
   time=rep(c("Hour2", "Hour10"), each=3),
   rep=paste0("Rep", 1:3))
# default will sort each column alphanumerically
pasteByRowOrdered(df1)
#>                   1                   2                   3                   4 
#>  Control_Hour2_Rep1  Control_Hour2_Rep2  Control_Hour2_Rep3 Control_Hour10_Rep1 
#>                   5                   6                   7                   8 
#> Control_Hour10_Rep2 Control_Hour10_Rep3     ABC1_Hour2_Rep1     ABC1_Hour2_Rep2 
#>                   9                  10                  11                  12 
#>     ABC1_Hour2_Rep3    ABC1_Hour10_Rep1    ABC1_Hour10_Rep2    ABC1_Hour10_Rep3 
#> 12 Levels: ABC1_Hour2_Rep1 ABC1_Hour2_Rep2 ABC1_Hour2_Rep3 ... Control_Hour10_Rep3

# keepOrder=TRUE will honor existing order of character columns
pasteByRowOrdered(df1, keepOrder=TRUE)
#>                   1                   2                   3                   4 
#>  Control_Hour2_Rep1  Control_Hour2_Rep2  Control_Hour2_Rep3 Control_Hour10_Rep1 
#>                   5                   6                   7                   8 
#> Control_Hour10_Rep2 Control_Hour10_Rep3     ABC1_Hour2_Rep1     ABC1_Hour2_Rep2 
#>                   9                  10                  11                  12 
#>     ABC1_Hour2_Rep3    ABC1_Hour10_Rep1    ABC1_Hour10_Rep2    ABC1_Hour10_Rep3 
#> 12 Levels: Control_Hour2_Rep1 Control_Hour2_Rep2 ... ABC1_Hour10_Rep3