Using multiple if statement, between conditions, inside a for loop

Question

I am a new R user and I have been trying to execute an if statement nested within a for loop in order to code a new variable. I have a data.frame where some guys previously forgot to code the "condition" variable (factor with 3 levels: old,new,lure) from E-prime. The task has two phases encoding/retrieval (Block 1 and 2), two set of images (A and B) and a unique Word ID.

So I have this:

phase <- rep(c("Block1", "Block2"), each = 7)
condition <- rep(NA, times = 14)
setAorB <- rep(c("A", "B"), times = c(9,5)) 
WordID <- c(23,34,56,76,45,88,99,23,34,56,76,45,100,105)

loris_data <- data.frame(phase,condition,setAorB,WordID)

which gives me:

> loris_data
   phase     condition setAorB WordID
1  Block1        NA       A     23
2  Block1        NA       A     34
3  Block1        NA       A     56
4  Block1        NA       A     76
5  Block1        NA       A     45
6  Block1        NA       A     88
7  Block1        NA       A     99
8  Block2        NA       A     23
9  Block2        NA       A     34
10 Block2        NA       B     56
11 Block2        NA       B     76
12 Block2        NA       B     45
13 Block2        NA       B    100
14 Block2        NA       B    105

What I would like to achieve is: At retrieval (Block2), if setAorB is "A", then condition is "old". I tried this basic loop but, obviously, worked only for the old items given that it does not discriminate lures vs new items.

for(i in 1:length(loris_data$condition)) {
      if(loris_data$setAorB[i] == "A") {
            loris_data$condition[i] <-"old"}
      else {
            loris_data$condition[i] <- "new"
      }
    }

Then, I would like to say: if setAorB is "B" and the Word ID is the same of A (which means that are lures), then the condition is "lure", otherwise if setAorB is "B" but it has a unique WordID, the the condition is "new".

This would be the expected output:

> loris_data
    phase    condition setAorB WordID
1  Block1      <NA>       A     23
2  Block1      <NA>       A     34
3  Block1      <NA>       A     56
4  Block1      <NA>       A     76
5  Block1      <NA>       A     45
6  Block1      <NA>       A     88
7  Block1      <NA>       A     99
8  Block2       old       A     23
9  Block2       old       A     34
10 Block2      lure       B     56
11 Block2      lure       B     76
12 Block2      lure       B     45
13 Block2       new       B    100
14 Block2       new       B    105

Can anyone help with this code as I am still learning and I am struggling quite a lot?

I guess it's easy to do achieve what you want, but please post expected final output. — pogibas
– pogibas, Commented Nov 11, 2017 at 14:06
I have just edited the post including the expected final output — Lollo
– Lollo, Commented Nov 11, 2017 at 14:23
Perhaps you need library(data.table); setDT(loris_data)[phase == "Block2", condition := c('new', 'old', 'lure')[as.integer(factor(1 + 2*(setAorB == "A") + 4 * (setAorB == "B" & WordID %in% loris_data$WordID[loris_data$setAorB=="A"])))] ] convert the column condition to character or use condition <- rep(NA_character_, times = 14) — akrun
– akrun, Commented Nov 11, 2017 at 14:29

pogibas · Accepted Answer · 2017-11-11 14:30:22Z

1

Quick and dirty solution using data.table:

library(data.table)
setDT(loris_data)
loris_data[, condition := ifelse(setAorB == "A", "old", "new")]
loris_data[phase != "Block2", condition := NA]
loris_data[phase == "Block2" & setAorB == "B" & WordID %in% loris_data[phase == "Block1", WordID], condition := "lure"]

answered Nov 11, 2017 at 14:30

pogibas

28.5k21 gold badges92 silver badges120 bronze badges

Sign up to request clarification or add additional context in comments.

3 Comments

pogibas Over a year ago

@LorisNaspi Happy to help :-)

Lollo Over a year ago

Would you suggest data.tables over data.frames?

pogibas Over a year ago

@LorisNaspi if your data is big, then I would go for data.table

www · Accepted Answer · 2017-11-11 15:22:48Z

A solution using dplyr. loris_data2 is the final output.

library(dplyr)

loris_data2 <- loris_data %>%
  group_by(WordID) %>%
  mutate(WordID_count = row_number()) %>%
  ungroup() %>%
  mutate(condition = case_when(
    phase %in% "Block2" & setAorB %in% "A"                        ~ "old",
    phase %in% "Block2" & setAorB %in% "B" & WordID_count > 1     ~ "lure",
    phase %in% "Block2" & setAorB %in% "B" & WordID_count == 1    ~ "new",
    TRUE                                                          ~ NA_character_
  )) %>%
  select(-WordID_count)

loris_data2
# # A tibble: 14 x 4
#     phase condition setAorB WordID
#    <fctr>     <chr>  <fctr>  <dbl>
#  1 Block1      <NA>       A     23
#  2 Block1      <NA>       A     34
#  3 Block1      <NA>       A     56
#  4 Block1      <NA>       A     76
#  5 Block1      <NA>       A     45
#  6 Block1      <NA>       A     88
#  7 Block1      <NA>       A     99
#  8 Block2       old       A     23
#  9 Block2       old       A     34
# 10 Block2      lure       B     56
# 11 Block2      lure       B     76
# 12 Block2      lure       B     45
# 13 Block2       new       B    100
# 14 Block2       new       B    105

Explanation

My solution first creates a new column called WordID_count, which shows the appearance times of a WordID. This task is achieved by the following.

loris_data %>%
  group_by(WordID) %>%
  mutate(WordID_count = row_number()) %>%
  ungroup()

# # A tibble: 14 x 5
#     phase condition setAorB WordID WordID_count
#    <fctr>     <lgl>  <fctr>  <dbl>        <int>
#  1 Block1        NA       A     23            1
#  2 Block1        NA       A     34            1
#  3 Block1        NA       A     56            1
#  4 Block1        NA       A     76            1
#  5 Block1        NA       A     45            1
#  6 Block1        NA       A     88            1
#  7 Block1        NA       A     99            1
#  8 Block2        NA       A     23            2
#  9 Block2        NA       A     34            2
# 10 Block2        NA       B     56            2
# 11 Block2        NA       B     76            2
# 12 Block2        NA       B     45            2
# 13 Block2        NA       B    100            1
# 14 Block2        NA       B    105            1

After that, it is ready to fill in the condition column. This task is achieved by the following.

mutate(condition = case_when(
    phase %in% "Block2" & setAorB %in% "A"                        ~ "old",
    phase %in% "Block2" & setAorB %in% "B" & WordID_count > 1     ~ "lure",
    phase %in% "Block2" & setAorB %in% "B" & WordID_count == 1    ~ "new",
    TRUE                                                          ~ NA_character_
  ))

mutate is the function to create a or update a new column. case_when is an alternative to multiple ifelse statement. The code did the following work:

if phase matches Block2 and setAorB matches A, the conditon is old.
if phase matches Block2 and setAorB matches A and WordID_count is larger than 1, the conditon is lure.
if phase matches Block2 and setAorB matches A and WordID_count is larger is 1, the conditon is new.
if none of the above situations, the conditon should be NA.

The final piece of the code is select(-WordID_count), which simply removes the WordID_count column as it is not part of the original data frame.

Santi · Accepted Answer · 2017-11-11 14:16:56Z

0

Have you tried an ifelse approach? ifelse is a built-in function in R that runs a vectorised if test and gives a determinate output. For example:

loris_data$resutl <- ifelse(test = loris_data[, "condition"] == "A",
       yes =  "old", 
       no = "new")

However, if you would like to nest another ifelse inside the no, that would be perfectly fine! Let me know if it works.

answered Nov 11, 2017 at 14:16

Santi

3682 silver badges16 bronze badges

1 Comment

Lollo Over a year ago

loris_data$condition <- ifelse(test = loris_data[, "setAorB"] == "A", yes = "old", no = "new") This code does not discriminate between "lure" and "new" item, tough. Please, look my expected output

Collectives™ on Stack Overflow

Using multiple if statement, between conditions, inside a for loop

3 Answers 3

3 Comments

1 Comment

1 Comment

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

3 Answers 3

3 Comments

1 Comment

1 Comment

Your Answer

Sign up or log in

Post as a guest

Related