How can I check if a value in one column is NA and if another is not NA?

Question

I have a data frame with three columns:

structure(list(introduction = structure(c(1522175035.901, 1520961618.904, 
1520978326.696, 1520965207.553, 1523918791.109, 1522173494.567, 
1521218841.268, 1522170731.4, 1521042736.772, 1522701901.544, 
1521145185.29, 1522706004.953, 1521150062.849, 1521047868.109, 
1520974602.5, 1521163554.751, 1524613269.765, 1521592575.346, 
1521061958.284, 1523045730.56, 1521219389.032, 1521828244.272, 
1521996291.152, 1522067323.354, 1525645608.604, 1521567579.468, 
1521567534.356, 1522426237.303, 1521567542.767, 1521567641.958, 
1521567555.865, 1521573065.994, 1521584733.425, 1522103377.939, 
1521821245.259, 1521829777.22, 1521829835.529, 1521980738.646, 
1522170449.584, 1522103997.818, 1522104390.669, 1522165861.644, 
1522171832.317, 1522171870.626, 1522171875.693, 1522171866.211, 
1522171933.098, 1522171855.451, 1522180873.26, 1522181644.842, 
1522333570.248, 1523525628.687, 1522487903.13, 1522078502.859, 
1522318895.073, 1522319216.787, 1522335495.615, 1522342148.729, 
1522342669.515, 1522692396.853, 1522832258.045, 1522860054.396, 
1523908701.209, 1522942128.02, 1523562329.681, 1523562640.799, 
1523569245.794, 1524162373.274, 1524599570.324, 1524249914.822, 
1524256541.826, 1525558540.165, 1524266810.439, 1525620261.23, 
1524520384.02, 1524680108.876, 1524678310.567, 1524768468.141, 
1524769276.356, 1524774051.792, 1524853389.661, 1525620789.277, 
1525626502.563, 1525649018.551, 1526072632.14, 1526484415.769, 
1526413818.926, 1526416221.779, 1527373743.794, 1527200067.957, 
1527642278.236, 1527719425.739, 1528335439.152), class = c("POSIXct", 
"POSIXt")), demoChat = structure(c(NA, NA, NA, NA, NA, NA, NA, 
NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, 1521996371.505, 
1522089543.872, NA, 1521567907.885, 1521567794.224, NA, 1521567970.389, 
1521568055.997, 1521567858.716, NA, NA, NA, 1521821640.607, 1521830168.929, 
1521830168.182, 1521980950.771, NA, 1522104226.885, 1522104740.322, 
NA, 1522172233.515, 1522172208.375, 1522172220.053, 1522172210.93, 
1522172211.298, 1522172241.241, 1522338283.99, NA, NA, 1523526243.697, 
1522488263.46, NA, 1522318901.563, NA, 1522335710.692, 1522342944.839, 
NA, 1522692960.938, NA, 1522860380.603, NA, 1522942437.122, NA, 
NA, 1523569647.089, NA, NA, 1524503493.328, NA, 1525668259.52, 
1524266917.649, 1525620673.917, 1526306725.36, NA, NA, 1524769206.836, 
1524769284.502, 1524774356.781, 1524853879.681, NA, NA, 1525649108.505, 
NA, 1526484828.946, 1526413874.213, 1526416344.729, 1528407164.373, 
1527200222.721, 1527642486.482, 1527719632.36, NA), class = c("POSIXct", 
"POSIXt")), demoChatSkipped = structure(c(1522175035.901, 1520961618.904, 
1520978326.696, 1520965207.553, 1523918791.109, 1522173494.567, 
1521218841.268, 1522170731.4, 1521042736.772, 1522701901.544, 
1521145185.29, 1522706004.953, 1521150062.849, 1521047868.109, 
1520974602.5, 1521163554.751, 1524613269.765, 1521592575.346, 
1521061958.284, 1523045730.56, 1521219389.032, 1521828244.272, 
NA, NA, NA, NA, NA, 1522426243.108, NA, NA, NA, 1521573070.218, 
1523474984.126, 1522103382.51, NA, NA, NA, NA, 1522170464.136, 
NA, NA, 1522165866.172, NA, NA, NA, NA, NA, NA, 1522180954.076, 
1522181651.368, 1522333575.922, NA, NA, 1522078502.859, NA, 1522319239.67, 
NA, NA, 1522342677.026, NA, 1522832258.045, NA, 1523908713.755, 
NA, 1523562337.321, 1523562650.926, NA, 1524162384.352, 1524599577.553, 
NA, 1524256546.09, 1525652931.413, NA, NA, 1524520476.233, NA, 
1524678321.292, NA, NA, NA, NA, 1525620799.957, 1525626519.183, 
NA, 1526072640.635, NA, NA, NA, 1527373758.872, NA, NA, NA, 1528335444.431
), class = c("POSIXct", "POSIXt"))), row.names = c(NA, -93L), class = "data.frame")

Using only the second and third columns, I want to count how many rows have an instance of is.na(demoChat) and !is.na(demoChat) and vice versa.

To make it clearer, I want to know (not in R format):

1. demoChat != NA && demoChatSkipped == NA
2. demoChat == NA && demoChatSkipped != NA

In R format:

!is.na(demoChat) && is.na(demoChatSkipped)
is.na(demoChat) && !is.na(demoChatSkipped)

I expect the best way to do this is to use table(); however, when I test for both values, I don't get the results I expect to get, which is a table of TRUE and FALSE values in relation to each statement.

How is this done?

Answer 1

You can use rowSums on a logical matrix of the conditions you want and then table the results.

table(rowSums(cbind(is.na(df[[2]]), !is.na(df[[3]]))))
# 0  1  2 
#42  6 45

The ones you want are the values corresponding to the entries 0 and 2:

• There are 45 rows with both is.na(df[[2]]) and !is.na(df[[3]]);
• There are 42 rows with both !is.na(df[[2]]) and is.na(df[[3]]).

Answer 2

A dplyr solution:

library(dplyr)

df %>
 filter(is.na(demoChat) & !is.na(demoChatSkipped)) %>% 
 tally()

Or even:

df %>%
  tally(is.na(demoChat) & !is.na(demoChatSkipped))

---

   n
1 45

Answer 3

I like to run an ifelse loop when i am trying to create a new categorical variable with several inputs. for your data (assuming you save it as x) you could do this:

library(tidyverse)                                                                                                                                                                                                      ), class = c("POSIXct", "POSIXt"))), row.names = c(NA, -93L), class = "data.frame")
mutate(x, NA_check = ifelse(is.na(demoChat)==TRUE & is.na(demoChatSkipped)==FALSE, 'demo NA, Chat not NA',
                             ifelse(is.na(demoChat)==FALSE & is.na(demoChatSkipped)==TRUE, 'demo not NA, Chat NA', 'some other thing happend or more ifelse loops')))

table(x$NA_check) 

                         demo NA, Chat not NA 
                                           45 
                         demo not NA, Chat NA 
                                           42 
some other thing happend or more ifelse loops 
                                        6