A small task I often face in R is looking up values in one set of data to add to another data frame. For example, I have a set of results from analyses in a data frame, but the full labels I want to use in tables and figures separately:

results <- data.frame(
  disease = c("all_cause_mortality",
              "ihd",
              "is",
              "ich",
              "mve"),
  result = rbinom(5, 100, 0.2)
)
results

##               disease result
## 1 all_cause_mortality     23
## 2                 ihd     21
## 3                  is     24
## 4                 ich     17
## 5                 mve     20

labels <- data.frame(
  disease = c("all_cause_mortality",
              "ihd",
              "is",
              "ich",
              "mve"),
  label = c("All-cause mortality",
            "Ischaemic heart disease",
            "Ischaemic stroke",
            "Intracerebral haemorrhage",
            "Major vascular event")
)
labels

##               disease                     label
## 1 all_cause_mortality       All-cause mortality
## 2                 ihd   Ischaemic heart disease
## 3                  is          Ischaemic stroke
## 4                 ich Intracerebral haemorrhage
## 5                 mve      Major vascular event

Here are a few ways to add the labels to the results…

1 Merge or join

merge() or dplyr::left_join() can be used to merge the two data frames:

merge(results, labels, by = "disease", sort = FALSE)

##               disease result                     label
## 1 all_cause_mortality     23       All-cause mortality
## 2                 ihd     21   Ischaemic heart disease
## 3                  is     24          Ischaemic stroke
## 4                 ich     17 Intracerebral haemorrhage
## 5                 mve     20      Major vascular event

dplyr::left_join(results, labels, by = "disease")

##               disease result                     label
## 1 all_cause_mortality     23       All-cause mortality
## 2                 ihd     21   Ischaemic heart disease
## 3                  is     24          Ischaemic stroke
## 4                 ich     17 Intracerebral haemorrhage
## 5                 mve     20      Major vascular event

The data frames can be merged the other way around to use the labels data to subset or choose the order:

labels2 <- data.frame(
  disease = c("mve",
              "is",
              "ich"),
  label = c("Major vascular event",
            "Ischaemic stroke",
            "Intracerebral haemorrhage")
)

merge(labels2, results, by = "disease", sort = FALSE)

##   disease                     label result
## 1     mve      Major vascular event     20
## 2      is          Ischaemic stroke     24
## 3     ich Intracerebral haemorrhage     17

dplyr::left_join(labels2, results, by = "disease")

##   disease                     label result
## 1     mve      Major vascular event     20
## 2      is          Ischaemic stroke     24
## 3     ich Intracerebral haemorrhage     17

This approach is also particularly useful when multiple columns are needed to join the data frames.

2 Match

match() can be used to identify the first appearance of disease and extract the correct label:

table <- results
table$disease <- labels$label[match(table$disease, labels$disease)]
table

##                     disease result
## 1       All-cause mortality     23
## 2   Ischaemic heart disease     21
## 3          Ischaemic stroke     24
## 4 Intracerebral haemorrhage     17
## 5      Major vascular event     20

3 Named vector

Putting the labels in a named vector and subsetting using [] is a concise approach:

labels <- c(
  all_cause_mortality = "All-cause mortality",
  ihd= "Ischaemic heart disease",
  is = "Ischaemic stroke",
  ich = "Intracerebral haemorrhage",
  mve = "Major vascular event"
)

table <- results
table$disease <- labels[table$disease]
table

##                     disease result
## 1       All-cause mortality     23
## 2   Ischaemic heart disease     21
## 3          Ischaemic stroke     24
## 4 Intracerebral haemorrhage     17
## 5      Major vascular event     20

Use match() to also subset and order:

labels2 <- c(
  mve = "Major vascular event",
  is = "Ischaemic stroke",
  ich = "Intracerebral haemorrhage"
)

table <- results
table$disease <- labels2[table$disease]
table[match(labels2, table$disease),]

##                     disease result
## 5      Major vascular event     20
## 3          Ischaemic stroke     24
## 4 Intracerebral haemorrhage     17

The named vector can also be useful in the labels argument of a discrete scale in a ggplot:

library(ggplot2)

ggplot(results) +
  geom_col(aes(result, disease),
           width = 0.5) +
  scale_y_discrete(limits = rev(names(labels)),
                   labels = labels) +
  theme_classic() +
  theme(axis.ticks.y = element_blank(),
        axis.line.y = element_blank(),
        axis.title.y = element_blank())