library(joyn)
#>
#> Attaching package: 'joyn'
#> The following object is masked from 'package:base':
#>
#> merge
library(data.table)
x <- data.table(id = c(1, 4, 2, 3, NA),
t = c(1L, 2L, 1L, 2L, NA),
country = c(16, 12, 3, NA, 15))
y <- data.table(id = c(1, 2, 5, 6, 3),
gdp = c(11L, 15L, 20L, 13L, 10L),
country = 16:20)This vignette will let you explore some additional features available
in joyn, through an example use case.
Suppose you want to join tables x and y,
where the variable country is available in both. You could do
one of five things:
If you don’t use the argument by, joyn will
consider country and id as key variables by default
given that they are common between x and
y.
# The variables with the same name, `id` and `country`, are used as key
# variables.
joyn(x = x,
y = y)
#>
#> ── JOYn Report ──
#>
#> .joyn n percent
#> <char> <int> <char>
#> 1: x 4 44.4%
#> 2: x & y 1 11.1%
#> 3: y 4 44.4%
#> 4: total 9 100%
#> ────────────────────────────────────────────────────────── End of JOYn report ──
#> ℹ ❯ Joyn's report available in variable .joyn
#> ℹ ❯ Removing key variables id and country from id, gdp, and country
#> Key: <id, country>
#> id t country gdp .joyn
#> <num> <int> <num> <int> <char>
#> 1: NA NA 15 NA x
#> 2: 1 1 16 11 x & y
#> 3: 2 1 3 NA x
#> 4: 2 NA 17 15 y
#> 5: 3 2 NA NA x
#> 6: 3 NA 20 10 y
#> 7: 4 2 12 NA x
#> 8: 5 NA 18 20 y
#> 9: 6 NA 19 13 yAlternatively, you can specify to join by country
# Joining by country
joyn(x = x,
y = y,
by = "country")
#>
#> ── JOYn Report ──
#>
#> .joyn n percent
#> <char> <int> <char>
#> 1: x 4 44.4%
#> 2: x & y 1 11.1%
#> 3: y 4 44.4%
#> 4: total 9 100%
#> ────────────────────────────────────────────────────────── End of JOYn report ──
#> ℹ ❯ Joyn's report available in variable .joyn
#> ℹ ❯ Removing key variables country from id, gdp, and country
#> Key: <country>
#> id t country gdp .joyn
#> <num> <int> <num> <int> <char>
#> 1: 3 2 NA NA x
#> 2: 2 1 3 NA x
#> 3: 4 2 12 NA x
#> 4: NA NA 15 NA x
#> 5: 1 1 16 11 x & y
#> 6: NA NA 17 15 y
#> 7: NA NA 18 20 y
#> 8: NA NA 19 13 y
#> 9: NA NA 20 10 yy and
don’t bring it into the resulting tableThis the default if you did not include country as part of
the key variables in argument by.
joyn(x = x,
y = y,
by = "id")
#>
#> ── JOYn Report ──
#>
#> .joyn n percent
#> <char> <int> <char>
#> 1: x 2 28.6%
#> 2: x & y 3 42.9%
#> 3: y 2 28.6%
#> 4: total 7 100%
#> ────────────────────────────────────────────────────────── End of JOYn report ──
#> ℹ ❯ Joyn's report available in variable .joyn
#> ℹ ❯ Removing key variables id from id, gdp, and country
#> Key: <id>
#> id t country gdp .joyn
#> <num> <int> <num> <int> <char>
#> 1: NA NA 15 NA x
#> 2: 1 1 16 11 x & y
#> 3: 2 1 3 15 x & y
#> 4: 3 2 NA 10 x & y
#> 5: 4 2 12 NA x
#> 6: 5 NA NA 20 y
#> 7: 6 NA NA 13 yAnother possibility is to make use of the update_NAs
argument of joyn(). This allows you to update the NAs
values in variable country in table x with the
actual values of the matching observations in country from
table y. In this case, actual values in country from table x
will remain unchanged.
joyn(x = x,
y = y,
by = "id",
update_NAs = TRUE)
#>
#> ── JOYn Report ──
#>
#> .joyn n percent
#> <char> <int> <char>
#> 1: NA updated 3 42.9%
#> 2: x 2 28.6%
#> 3: x & y 2 28.6%
#> 4: total 7 100%
#> ────────────────────────────────────────────────────────── End of JOYn report ──
#> ℹ ❯ Joyn's report available in variable .joyn
#> ℹ ❯ Removing key variables id from id, gdp, and country
#> Key: <id>
#> id t country gdp .joyn
#> <num> <int> <num> <int> <char>
#> 1: NA NA 15 NA x
#> 2: 1 1 16 11 x & y
#> 3: 2 1 3 15 x & y
#> 4: 3 2 20 10 NA updated
#> 5: 4 2 12 NA x
#> 6: 5 NA 18 20 NA updated
#> 7: 6 NA 19 13 NA updatedYou can also update all the values - both NAs and actual - in
variable country of table x with the actual values
of the matching observations in country from y.
This is done by setting update_values = TRUE.
Notice that the reportvar allows you keep track of how
the update worked. In this case, value update means that only
the values that are different between country from
x and country from y are updated.
However, let’s consider other possible cases:
If, for the same matching observations, the values between the two country variables were the same, the reporting variable would report x & y instead (so you know that there is no update to make).
if there are NAs in country from y, the
actual values in x will be unchanged, and you would see a
not updated status in the reporting variable. Nevertheless,
notice there is another way for you to bring country from
y to x. This is done through the argument
keep_y_in_x (see 5. below ⬇️)
# Notice that only the value that are
joyn(x = x,
y = y,
by = "id",
update_values = TRUE)
#>
#> ── JOYn Report ──
#>
#> .joyn n percent
#> <char> <int> <char>
#> 1: NA updated 3 42.9%
#> 2: not updated 2 28.6%
#> 3: value updated 2 28.6%
#> 4: total 7 100%
#> ────────────────────────────────────────────────────────── End of JOYn report ──
#> ℹ ❯ Joyn's report available in variable .joyn
#> ℹ ❯ Removing key variables id from id, gdp, and country
#> Key: <id>
#> id t country gdp .joyn
#> <num> <int> <num> <int> <char>
#> 1: NA NA 15 NA not updated
#> 2: 1 1 16 11 value updated
#> 3: 2 1 17 15 value updated
#> 4: 3 2 20 10 NA updated
#> 5: 4 2 12 NA not updated
#> 6: 5 NA 18 20 NA updated
#> 7: 6 NA 19 13 NA updatedAnother available option is that of bringing the original variable
country from y into the resulting table, without
using it to update the values in x. In order to distinguish
country from x and country from
y, joyn will assign a suffix to the variable’s
name: so that you will get country.y and country.x.
All of this can be done specifying
keep_common_vars = TRUE.
joyn(x = x,
y = y,
by = "id",
keep_common_vars = TRUE)
#>
#> ── JOYn Report ──
#>
#> .joyn n percent
#> <char> <int> <char>
#> 1: x 2 28.6%
#> 2: x & y 3 42.9%
#> 3: y 2 28.6%
#> 4: total 7 100%
#> ────────────────────────────────────────────────────────── End of JOYn report ──
#> ℹ ❯ Joyn's report available in variable .joyn
#> ℹ ❯ Removing key variables id from id, gdp, and country
#> Key: <id>
#> id t country.x gdp country.y .joyn
#> <num> <int> <num> <int> <int> <char>
#> 1: NA NA 15 NA NA x
#> 2: 1 1 16 11 16 x & y
#> 3: 2 1 3 15 17 x & y
#> 4: 3 2 NA 10 20 x & y
#> 5: 4 2 12 NA NA x
#> 6: 5 NA NA 20 18 y
#> 7: 6 NA NA 13 19 yIn joyn , you can also bring non common variables from
y into the resulting table. In fact you can specify them in
y_vars_to_keep, as shown in the example below:
# Keeping variable gdp
joyn(x = x,
y = y,
by = "id",
y_vars_to_keep = "gdp")
#>
#> ── JOYn Report ──
#>
#> .joyn n percent
#> <char> <int> <char>
#> 1: x 2 28.6%
#> 2: x & y 3 42.9%
#> 3: y 2 28.6%
#> 4: total 7 100%
#> ────────────────────────────────────────────────────────── End of JOYn report ──
#> ℹ ❯ Joyn's report available in variable .joyn
#> Key: <id>
#> id t country gdp .joyn
#> <num> <int> <num> <int> <char>
#> 1: NA NA 15 NA x
#> 2: 1 1 16 11 x & y
#> 3: 2 1 3 15 x & y
#> 4: 3 2 NA 10 x & y
#> 5: 4 2 12 NA x
#> 6: 5 NA NA 20 y
#> 7: 6 NA NA 13 yNotice that if you set y_vars_to_keep = FALSE or
y_vars_to_keep = NULL, then joyn won’t bring
any variable into the returning table.