# fmtr

The fmtr package helps format data. The package aims to simulate the basic functionality of SAS® formats, but with R. The package contains several functions that make formatting simpler and more powerful.

### Key Functions

fmtr contains the following key functions:

• The fdata() function to apply formatting to any data frame or tibble.
• The fapply() function to apply formatting to any vector.
• The formats() and fattr() functions to easily assign formatting attributes.
• The value() and condition() functions to create a user-defined format.
• The fcat() function to create a format catalog.
• The flist() function to create a formatting list.

### How to Use

The fmtr package builds heavily on existing R formatting capabilities. For most R programmers, these functions are well-known, and widely used. The examples below make use of standard R formatting codes, such as those associated with the strptime() and sprintf() functions. The standard R formatting codes are a flexible and compact way of defining a format. If you are unfamiliar with R formatting codes, please see this summary on the FormattingStrings() page.

#### The fapply() Function

The simplest way to introduce the fmtr package is to examine the use of the fapply() function.

library(fmtr)

# Create sample data vector
v1 <- c(1.483, 5.29837, 7.9472, 8.684021)

# Apply format
fapply(v1, "%1.1f")
# [1] "1.5" "5.3" "7.9" "8.7"

As you can see from the above example, the fapply() function typically takes two parameters: a vector and a format. In this way, the fapply() acts very much like a SAS® put function.

Note that the format parameter can also be assigned as an attribute on the vector. The fapply() function will then pick up the format attribute, and apply it to the input vector. The result is the same:

library(fmtr)

# Create sample data vector
v1 <- c(1.483, 5.29837, 7.9472, 8.684021)

# Assign format attribute
attr(v1, "format") <- "%1.1f"

# Apply format
fapply(v1)
# [1] "1.5" "5.3" "7.9" "8.7"

Besides the format attribute, the fapply() function will also recognize attributes for width and justify. These parameters allow you to control the width and alignment of the data in the vector. If the width parameter is larger than the width of the data, the value will be padded with spaces. Here is an example:

library(fmtr)

# Create sample data vector
v1 <- c(1.483, 5.29837, 7.9472, 8.684021)

# Assign formatting attributes
attr(v1, "format") <- "%1.1f"
attr(v1, "width") <- 5
attr(v1, "justify") <- "right"

# Apply formatting attributes
fapply(v1)
# [1] "  1.5" "  5.3" "  7.9" "  8.7"

To help simplify assignment of these attributes, the fmtr package includes the fattr() function, which allows you to set all the above attributes in one function call. Here is an example using the fattr() function, that ends with the same result as the example above.

library(fmtr)

# Create sample data vector
v1 <- c(1.483, 5.29837, 7.9472, 8.684021)

# Assign formatting attributes
v1 <- fattr(v1, format = "%1.1f", width = 5, justify = "right")

# Apply formatting attributes
fapply(v1)
# [1] "  1.5" "  5.3" "  7.9" "  8.7"

Note that fapply() can accept several different types of formats. The examples above focus on a simple numeric format. But fapply() also accepts date formats, a lookup list, a user-defined format, a vectorized function, and a formatting list.

Here is an example showing the use of a lookup list:

library(fmtr)

# Create sample data vector
v1 <- c("A", "B", "A", "C", "B")

# Create lookup vector
v2 <- c(A = "Group A", B = "Group B", C = "Group C")

fapply(v1, v2)
# [1] "Group A" "Group B" "Group A" "Group C" "Group B"

#### A User-Defined Format

The weakness with using a named vector as a lookup list, as in the above example, is that there is no way to include any sort of logic in the lookup. For instance, if your data has NA values, you may want to handle those differently from the valid input values. Or you may want to define a default value if the input data does not match any of the lookup keys.

For these reasons, the fmtr package provides a user-defined format. This concept was taken directly from SAS® software. The functions that create a user-defined format are value() and condition().

A condition accepts an expression and a label. The expression determines which label is assigned. For the expression, you can use logical operators like “&” and “|”, and relational operators like “>” and “<”. The data value is identified with a variable “x”. Here is an example:

library(fmtr)

# Create sample data vector
v1 <- c("A", "B", "E", "A", NA, "C", "D")

u1 <- value(condition(x == "A", "Group A"),
condition(x == "B", "Group B"),
condition(x == "C" | x == "D", "Group C/D"),
condition(TRUE, "Other"))

fapply(v1, u1)
# [1] "Group A" "Group B" "Other" "Group A" "Other" "Group C/D" "Group C/D"

Notice that the user-defined format gives you much more capabilities than a simple lookup vector. It allows you to perform categorization, and assign a default. Additionally, the NA missing value does not crash the function. The NA simply falls into the default category. If there is no default category, any values which do not correspond to a category will fall through the format unaltered.

#### The fdata() Function

The fdata() function works very much the same way as fapply(), but with data frames and tibbles instead of vectors. In fact, under the hood, fdata() is simply calling fapply() for each column in the data frame.

Like the fapply() function, formatting may be assigned to data frame columns using the format, width, and justify attributes. Formatting is then applied by calling the fdata() function, and passing the data frame as the first parameter. fdata() will then return a new data frame with the specified formatting applied. This method of formatting provides much greater control than the base R format() function.

library(fmtr)

# Construct data frame from state vectors
df <- data.frame(state = state.abb, area = state.area)[1:10, ]

# Calculate percentages
df$pct <- df$area / sum(state.area) * 100

# Before formatting
df
#    state   area         pct
# 1     AL  51609  1.42629378
# 2     AK 589757 16.29883824
# 3     AZ 113909  3.14804973
# 4     AR  53104  1.46761040
# 5     CA 158693  4.38572418
# 6     CO 104247  2.88102556
# 7     CT   5009  0.13843139
# 8     DE   2057  0.05684835
# 9     FL  58560  1.61839532
# 10    GA  58876  1.62712846

# Create state name lookup list
name_lookup <- state.name
names(name_lookup) <- state.abb

# Assign formats
formats(df) <- list(state = name_lookup,
area  = function(x) format(x, big.mark = ","),
pct   = "%.1f%%")

# Apply formats
fdata(df)
#          state    area   pct
# 1      Alabama  51,609  1.4%
# 3      Arizona 113,909  3.1%
# 4     Arkansas  53,104  1.5%
# 5   California 158,693  4.4%
# 7  Connecticut   5,009  0.1%
# 8     Delaware   2,057  0.1%
# 9      Florida  58,560  1.6%
# 10     Georgia  58,876  1.6%

In the above example, observe that the formats() function assigns the format attribute for multiple columns. This assignment is accomplished by sending a named list into the formats() function, where the names in the list correspond to the column names of the data frame. Also note the use of a lookup style format for the state names, and an anonymous vectorized format function for the state area.

#### The fcat() Function

One of the benefits of the above method of formatting is that the data frame attributes can be stored with the data frame, and reapplied in the future. But what if you want to apply the same set of formats to a different data frame?

That is where you need a format catalog.

The format catalog is a collection of formats that can be saved and reused. A format catalog is created with an fcat() function. To create a format catalog, you call the fcat() function, passing a set of name/format pairs. In this case, the name of the format is a generic format name. It does not have to correspond to a column name. You may name the formats anything you want. The formats can be accessed in the catalog using dollar sign (“$”) list notation. library(fmtr) # Construct data frame from state vectors df <- data.frame(state = state.abb, area = state.area)[1:10, ] # Calculate percentages df$pct <- df$area / sum(state.area) * 100 # Before formatting df # state area pct # 1 AL 51609 1.42629378 # 2 AK 589757 16.29883824 # 3 AZ 113909 3.14804973 # 4 AR 53104 1.46761040 # 5 CA 158693 4.38572418 # 6 CO 104247 2.88102556 # 7 CT 5009 0.13843139 # 8 DE 2057 0.05684835 # 9 FL 58560 1.61839532 # 10 GA 58876 1.62712846 # Create state name lookup list name_lookup <- state.name names(name_lookup) <- state.abb # Assign formats to format catalog cat1 <- fcat(state = name_lookup, area = function(x) format(x, big.mark = ","), pct = "%.1f%%") # Apply a format from the catalog using fapply fapply(df$pct, cat1\$pct)
# [1] "1.4%"  "16.3%" "3.1%"  "1.5%"  "4.4%"  "2.9%"  "0.1%"  "0.1%"  "1.6%"  "1.6%"

# Assign formats from the catalog to format attributes
formats(df) <- cat1

# Apply formats
fdata(df)
#          state    area   pct
# 1      Alabama  51,609  1.4%
# 3      Arizona 113,909  3.1%
# 4     Arkansas  53,104  1.5%
# 5   California 158,693  4.4%
# 10     Georgia  58,876  1.6%
In normal use, of course, the format catalog would likely be created in a separate script and saved to a file using the write.fcat() function. The format catalog can then be read by any number of programs using the read.fcat() function, and the formats in the catalog can be applied as needed to your data.