r loop/function to find matches from a list - r

Question

I have a list of sales people in three columns and I want to go down a list I have and:
a) Where their name appears in any of the three column
b) their name appears with a Trainee Sales person (these would be people whose name is not in the list)
ilist <- c("SP1","SP2","SP3","SP4","SP5")
df2 <-
data.frame(sales1 = c("SP5","SP5","SP4","SP3","SP2","SP1","SP3"),
sales2 = c("","SP4","SP1","SP1","SP5","SP3",""),
sales3 = c("","SP9","","SP6","","",""))
Output I'd expect something like below answer(though i'd take any output):
A B
SP1 3 1
SP2 1 0
SP3 3 1
SP4 1 1
SP5 3 1
I've tried creating a loop and a function but i cant seem to get them to work.
The aim after getting this to work is to make it part of a group_by so I can break it down by type and year
data %>%
group_by(type,year) %>%
your helpful answer here
edit:
select of the columns I'm looking at using.
My iList would be something like the below
(in the 3 Columns column 2 and 3 will contain blanks where the sales person only appears in column 1; there is also no set placement for where a sales person or trainee may appear)
ilist <- c("SJ","KW","MOLC","FERB","BACC")
structure(list(iYear = structure(c(1L, 4L, 3L, 4L, 4L,
4L, 5L, 5L, 6L, 9L), .Label = c("2020-07-01", "2020-07-02", "2020-07-03",
"2020-07-04", "2020-07-06", "2020-07-07", "2020-07-08", "2020-07-09",
"2020-07-10", "2020-07-11", "2020-07-12", "2020-07-13", "2020-07-14",
"2020-07-15", "2020-07-16", "2020-07-17", "2020-07-18", "2020-07-19",
"2020-07-20", "2020-07-21", "2020-07-22", "2020-07-23", "2020-07-24",
"2020-07-25", "2020-07-27", "2020-07-28", "2020-07-29", "2020-07-30",
"2020-07-31"), class = "factor"), iType = structure(c(4L,
4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L), .Label = c("", "ZB", "BS",
"CFN", "CTR", "MJ", "UK", "EFH", "ENOC", "EY", "F", "G", "CD",
"HAEM", "HN", "IC", "LB", "LY", "MNN", "MOS", "NERO", "ZZZ",
"ZZZQE", "GFT", "PG", "RE", "SK", "UR"), class = "factor"),
Sales.1 = structure(c(74L, 20L, 74L, 16L,
3L, 3L, 3L, 16L, 58L, 41L), .Label = c("", "ABUE", "AHMEM",
"AJOS", "ANNS", "AOK", "BACC", "BH", "BLAFM", "BLOCA", "BRAD",
"BROWNJ", "BRT", "BUIH", "BURDA", "BURYA", "CANRJ", "CAVM",
"CHAMBA", "COOSNP", "COUPSI", "CPH", "CTT", "DARA", "DILP",
"EXPAT", "FCH", "FERB", "FERMA", "GT", "GT", "HAEM", "HAMJR",
"HENJ", "HENJA", "HOWRA", "HUSA", "ILINC", "JONG", "KC",
"KNOT", "KW", "LAUC", "LOOP", "LYEJO", "LYNN", "MAJJ", "MCGREA",
"MENT", "MKB", "MOLC", "MUDHS", "MULLM", "NC", "NODS",
"O'BSG", "OLIT", "OLIVK", "PAEI", "PARKD", "PATEF", "PERT",
"POL", "PTRHUS", "RAMACN", "RAMS", "REYMA", "ROBCM", "ROBINE",
"SAMJN", "SAYC", "SHARMM", "SHEG", "SJ", "SJN", "SKINT",
"SLOP", "SORT", "SOUBIO", "SPOE", "TELED", "THAN", "THEL",
"TURH", "TURHJ", "UCONS", "UPH", "UT", "VALK", "WALJ"
), class = "factor"), Sales.2 = structure(c(1L,
12L, 1L, 1L, 1L, 1L, 1L, 1L, 3L, 45L), .Label = c("", "ABUE",
"AHMEM", "AJOS", "AOK", "BACC", "BH", "BLAFM", "BROWNJ",
"BUIH", "BURDA", "BURYA", "CANRJ", "CAVM", "CHAMBA", "COOSNP",
"COUPSI", "DARA", "DILP", "FCH", "FERB", "FERMA", "GYNT",
"HOWRA", "HUSA", "ILINC", "KW", "LAUC", "LOOP", "LYNN", "MAJJ",
"MOLC", "MULLM", "NC", "OLIVK", "PARKD", "POL", "PTRHUS",
"RAMS", "REYMA", "ROBCM", "ROBINE", "SAMJN", "SHARMM", "SJ",
"SJN", "SKINT", "SLOP", "SORT", "SPOE", "TELED", "THAN",
"THEL", "TURH", "VALK"), class = "factor"), Sales.3 = structure(c(1L,
1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L), .Label = c("", "AHMEM",
"AOK", "BACC", "BLAFM", "CHAMBA", "COUPSI", "DILP", "FCH",
"KW", "LOOP", "MAJJ", "PTRHUS", "RAMS", "ROBCM", "SAMJN",
"SHARMM", "SJ", "TELED", "THAN", "VALK"), class = "factor")), row.names = c(NA,
10L), class = "data.frame")

Answer 1

I'm not sure this is what you're looking for, but thought it might be helpful. With your interest in using group_by sounds like you might want a tidyverse approach.
Here, would add row numbers, so you can group_by each row to see if a trainee is present with a sales person in the same row.
Then, use pivot_longer to put into a long format, and remove the empty strings.
When grouping by the row number, you can add an indicator that those people will have appeared with a trainee sales person. It looks to see if the person is not contained in the ilist.
Finally, you can group_by each sales person, only include those in ilist with filter, and add up the number of times appeared (assumes only once per row in the initial data), and number of trainee contacts.
library(tidyverse)
df2 %>%
mutate(rn = row_number()) %>%
pivot_longer(cols = -rn) %>%
na_if("") %>%
na.omit %>%
group_by(rn) %>%
mutate(with_trainee = ifelse(any(!value %in% ilist), 1, 0)) %>%
group_by(value) %>%
filter(value %in% ilist) %>%
summarise(A = n(),
B = sum(with_trainee))
Output
value A B
<chr> <int> <dbl>
1 SP1 3 1
2 SP2 1 0
3 SP3 3 1
4 SP4 2 1
5 SP5 3 1
Edit 1: With your "live data", and grouping results by year from iYear and iType, you can try this:
library(tidyverse)
df2 %>%
mutate(rn = row_number(),
iYear = substr(iYear, 1, 4)) %>%
pivot_longer(cols = -c(rn, iYear, iType)) %>%
na_if("") %>%
na.omit %>%
group_by(rn, iYear, iType) %>%
mutate(with_trainee = ifelse(any(!value %in% ilist), 1, 0)) %>%
group_by(value, iYear, iType) %>%
filter(value %in% ilist) %>%
summarise(A = n(),
B = sum(with_trainee))
Edit 2: Additional detailed explanation:
Row numbers (rn via row_number) is helpful in this case, as you want to know if sales people are present at the same time (which implies "within the same row"). So, if 2 sales people share the same rn, they were present at the same time.
iYear is changed to just a year. It uses substr() (substring) to take the 1st through 4th character of iYear which in XXXX-XX-XX date format is the year.
pivot_longer (and its friend, pivot_wider) are really powerful to convert from long <-> wide formats of data. In the tidyr package, pivot_longer takes all the columns (except for rn, iYear, and iType) and puts into two columns (name and value). value contains the salesperson in a single column now, instead of multiple columns it started with.
na_if("") will make the blank strings "" become NA (missing data). The follow up na.omit will remove those rows with NA then.
The group_by with rn makes sure you are looking collectively at those sales people who share the same rn. I added iYear and iType so that they will also show up in the final summarized results. Then, with_trainee is a new column that will contain whether that sales person is with a trainee or not (after group_by using any to see if "any row" within the group, sharing the same rn, is in the ilist vector). If there is, then code as 1, if not, then code as 0.
The next group_by is by value (or sales person), using filter since you will only want results on those who are in ilist. (If you wanted everyone, including trainees not in ilist, you could leave this line out.)
The final summarise works with the group_by - with n() showing the number of rows of data per value (or per sales person), which is the same as number of different rn values a sales person may appear in overall. The sum(with_trainee) is the total number of times the with_trainee was 1 for a given value (or sales person).
Output
value iYear iType A B
<fct> <chr> <fct> <int> <dbl>
1 SJ 2020 CFN 3 1

Answer 2

I don't really understand the expected result to be honest since you say you expect SP2 | 1 | 0 but SP2 did not appear in line 1. the following might do what you want... or not.
library(data.table)
sales <- data.table(sale = c("SP1", "SP2", "SP3", "SP4", "SP5"))
sales_group <-
data.table(
sales1 = c("SP5", "SP5", "SP4", "SP3", "SP2", "SP1", "SP3"),
sales2 = c("", "SP4", "SP1", "SP1", "SP5", "SP3", ""),
sales3 = c("", "SP9", "", "SP6", "", "", "")
)
all <- sort(sales_group[, unique(c(sales1, sales2, sales3))])
all <- all[all != ""]
trainees <- all[!all %in% c(sales$sale, "")]
sales_group[, pos := seq(.N)]
sales1 <- merge(sales, sales_group, by.x = "sale", by.y = "sales1")
sales2 <- merge(sales, sales_group, by.x = "sale", by.y = "sales2")
sales3 <- merge(sales, sales_group, by.x = "sale", by.y = "sales3")
setnames(sales1, c("sale", "plusone", "plustwo", "sales_pos"))
setnames(sales2, c("sale", "plusone", "plustwo", "sales_pos"))
setnames(sales3, c("sale", "plusone", "plustwo", "sales_pos"))
sales_visit_by_sale <- rbind(sales1, sales2, sales3)
sales_visit_by_sale[, with_trainee := FALSE]
sales_visit_by_sale[(plusone %in% trainees) | (plustwo %in% trainees), with_trainee := TRUE]
sales_visit_by_sale[(order(sale, sales_pos)), .(sale, sales_pos, with_trainee)]