Looking up values from another dataframe in r

Question

I have large dataframe called df with some ID's.

I have another dataframe (id_list) with a set of matching ID's and its associated features for each ID. The ID are not sequentally ordered in both dataframes.

Effectively i would like to look up from the larger dataframe df to the id_list and add two columns namely Display and Type to the current dataframe df.

There are numerous confusing examples. What could be the most effective way of doing this. I tried using match() , %in% and failed miserably.

Here is a reproducible example.

df <- data.frame(Feats = matrix(rnorm(20), nrow = 20, ncol = 5), ID = sample.int(10, 10))

id_list <- data.frame(ID = sample.int(10,10),
           Display = sample(c('clear', 'blur'), 20, replace = TRUE),
           Type = sample(c('red', 'green', 'blue', 'indigo', 'yellow'), 20, replace = TRUE))

           Feats.1     Feats.2     Feats.3     Feats.4     Feats.5 ID
1   3.14944573 -0.52285062  3.14944573 -0.52285062  3.14944573  2
2  -0.41096007  0.38256691 -0.41096007  0.38256691 -0.41096007  1
3   0.03629351 -0.02514005  0.03629351 -0.02514005  0.03629351  7
4   0.91257290  1.35590761  0.91257290  1.35590761  0.91257290  5
5  -0.26927311 -2.10213773 -0.26927311 -2.10213773 -0.26927311  3
6   3.14944573 -0.52285062  3.14944573 -0.52285062  3.14944573  4
7  -0.41096007  0.38256691 -0.41096007  0.38256691 -0.41096007 10
8   0.03629351 -0.02514005  0.03629351 -0.02514005  0.03629351  6
9   0.91257290  1.35590761  0.91257290  1.35590761  0.91257290  8
10 -0.26927311 -2.10213773 -0.26927311 -2.10213773 -0.26927311  9

  ID Display   Type
1   6   clear indigo
2   1    blur   blue
3   7   clear    red
4   4   clear    red
5   3    blur    red
6  10   clear yellow
7   2   clear   blue
8   8    blur  green
9   5   clear   blue
10  9   clear  green

The resulting end df should be of size [20 x 8].

Answer 1

You can use merge from base R or left_join from dplyr to do this pretty easily. (There's also data.table::merge, which maybe someone else can give an answer with.) You probably want to take steps to ensure that you don't lose any data if there's an entry in your data frame that doesn't have a corresponding ID in the lookup. If that's not the case, you can change all.x to false or null in merge, or switch from left_join to inner_join. To illustrate, I added a dummy row to the data with an ID that doesn't exist in the lookup table.

df <- data.frame(Feats = matrix(rnorm(10), nrow = 5, ncol = 5), ID = sample.int(10, 10))
dummy <- df[1, ]
dummy$ID <- 12
df <- rbind(dummy, df)

id_list <- data.frame(ID = sample.int(10,10),
                      Display = sample(c('clear', 'blur'), 10, replace = TRUE),
                      Type = sample(c('red', 'green', 'blue', 'indigo', 'yellow'), 10, replace = TRUE))

With merge, you set either by as the column name from both data frames to join by, or by.x and by.y if they have different names. all.x = T will keep all observations in the first data frame even if they don't match an observation in the second data frame.

merged1 <- merge(df, id_list, by = "ID", sort = F, all.x = T)
merged1
#>    ID     Feats.1    Feats.2     Feats.3    Feats.4     Feats.5 Display
#> 1  10 -1.44053344  1.0086988 -1.44053344  1.0086988 -1.44053344   clear
#> 2   5  0.99220217 -0.3125813  0.99220217 -0.3125813  0.99220217   clear
#> 3   2  1.03881289  1.1277627  1.03881289  1.1277627  1.03881289   clear
#> 4   7 -0.01678186 -0.1519029 -0.01678186 -0.1519029 -0.01678186   clear
#> 5   4  0.07130125  1.1715833  0.07130125  1.1715833  0.07130125   clear
#> 6   6 -1.44053344  1.0086988 -1.44053344  1.0086988 -1.44053344   clear
#> 7   8  0.99220217 -0.3125813  0.99220217 -0.3125813  0.99220217    blur
#> 8   3  1.03881289  1.1277627  1.03881289  1.1277627  1.03881289   clear
#> 9   1 -0.01678186 -0.1519029 -0.01678186 -0.1519029 -0.01678186   clear
#> 10  9  0.07130125  1.1715833  0.07130125  1.1715833  0.07130125   clear
#> 11 12 -1.44053344  1.0086988 -1.44053344  1.0086988 -1.44053344    <NA>
#>      Type
#> 1  indigo
#> 2  yellow
#> 3    blue
#> 4  indigo
#> 5  yellow
#> 6  indigo
#> 7   green
#> 8     red
#> 9     red
#> 10   blue
#> 11   <NA>

dplyr::left_join keeps all observations from the first data frame and merges in any matching ones from the second.

joined <- dplyr::left_join(df, id_list, by = "ID")
head(joined)
#>       Feats.1    Feats.2     Feats.3    Feats.4     Feats.5 ID Display
#> 1 -1.44053344  1.0086988 -1.44053344  1.0086988 -1.44053344 12    <NA>
#> 2 -1.44053344  1.0086988 -1.44053344  1.0086988 -1.44053344 10   clear
#> 3  0.99220217 -0.3125813  0.99220217 -0.3125813  0.99220217  5   clear
#> 4  1.03881289  1.1277627  1.03881289  1.1277627  1.03881289  2   clear
#> 5 -0.01678186 -0.1519029 -0.01678186 -0.1519029 -0.01678186  7   clear
#> 6  0.07130125  1.1715833  0.07130125  1.1715833  0.07130125  4   clear
#>     Type
#> 1   <NA>
#> 2 indigo
#> 3 yellow
#> 4   blue
#> 5 indigo
#> 6 yellow

Created on 2018-07-13 by the reprex package (v0.2.0).