Assigning values according to set limits in R - r

Question

I have a list of stores with some quantities of different articles in them and a warehouse with these articles - these are two separate data frames.
Article <- c('a','b','a','b','c','d')
forecast <- c( 1,5,80,10,100,1000)
StoreID <- c(1,1,2,2,3,4)
StoreData <- data.frame(StoreID, Article, Order)
Smth like this:
StoreData
StoreID Article forecast
1 a 1
1 b 5
2 a 80
2 b 10
3 c 100
4 d 1000
And the warehouse data:
Stock <- c(10,11,12,100)
WarehouseData <- data.frame(Article, Stock)
WarehouseData
Article Stock
a 10
b 11
c 12
d 100
My target is to have a purchase order column. The logic has to be following: for every row in the StoreData table I have to look at the Stock of the Article in the Warehouse and if it is enough - approve the fcst, if not - approve only the avaulable quantity. My problem is that while approving quantities the avaialble stock is reducing and I cannot find out how to take it into account.
The expected result looks like this:
StoreData
StoreID Article forecast PO
1 a 1 1
1 b 5 5
2 a 80 9
2 b 10 6
3 c 100 12
4 d 1000 100
Can anyone, please, tell how to get it right?

Answer 1

Here's another approach using dplyr:
library(dplyr)
left_join(storeData, WarehouseData, by = "Article") %>%
group_by(Article) %>%
mutate(PO = ifelse(cumsum(forecast) <= Stock, forecast,
Stock - cumsum(forecast) + forecast)) %>% ungroup
#Source: local data frame [6 x 5]
#
# StoreID Article forecast Stock PO
# (int) (fctr) (int) (dbl) (dbl)
#1 1 a 1 10 1
#2 1 b 5 11 5
#3 2 a 80 10 9
#4 2 b 10 11 6
#5 3 c 100 12 12
#6 4 d 1000 100 100

Answer 2

See the loop below for example:
StoreData$PO <- NA
for (i in 1:nrow(StoreData)) {
query <- WarehouseData$Article == StoreData[i, "Article"]
po <- ifelse(StoreData[i, "forecast"] > WarehouseData[query, 2],
WarehouseData[query, 2],
StoreData[i, "forecast"])
WarehouseData[query, 2] <- WarehouseData[query, 2] - po
StoreData[i, "PO"] <- po
}
print(StoreData)
# StoreID Article forecast PO
# 1 1 a 1 1
# 2 1 b 5 5
# 3 2 a 80 9
# 4 2 b 10 6
# 5 3 c 100 12
# 6 4 d 1000 100
This is another alternative based on the other solution using base R:
StoreData <- merge(StoreData, WarehouseData)
StoreData$PO <- do.call(c, lapply(split(StoreData, StoreData$Article), function(z) {
ifelse(cumsum(z$forecast) <= z$Stock, z$forecast,
z$Stock - cumsum(z$forecast) + z$forecast)
}))
And here is what I used to recreate your dataset, might help other answers:
StoreData <- read.table(text = "StoreID Article forecast
1 a 1
1 b 5
2 a 80
2 b 10
3 c 100
4 d 1000", header = T)
Article <- c('a','b','c','d')
Stock <- c(10,11,12,100)
WarehouseData <- data.frame(Article, Stock)