R apply function to rows of data frame using values as arguments

r apply

10,230

Solution 1

With mapply you could do something like:

mapply(getDayRate, TRX$Date, TRX$Currency)

Solution 2

The output you want is not so clear. your function getDayRate retrieves directly, from the currencies table hard-coded into it, the ExchangeRate chosen date and currency which makes sense so far.

However, running your function against the same table it retrieves the values from we give the same result of just selecting that column...

> all.equal(mapply(getDayRate, currencies$Date, currencies$Currency), currencies$ExchangeRate, check.attributes=F)
[1] TRUE

Probably (I think) you actually want to do some joins (with the second table you provided?) linking another table to the currencies data.frame by date and currency.

EDIT:

your edit now clarifies things. why don't you try also this

library(reshape2)

melt(currencies)

10,230

user670186

Updated on September 16, 2022

Comments

user670186 over 1 year
I have a df TRX with pairs of Date and Curreny
```
Date            Currency      ExchangeRate
2012-08-13      EUR           ?
2012-08-13      CHF           ?
2012-08-13      CZK           ?
```
I have a second df CURRENCIES for the currency conversion rates with EUR base.
```
Date            EUR    CHF       CZK
2012-08-13      1      1.24      25.73
2012-08-13      1      1.23      25.92
2012-08-13      1      1.22      24.00
```
Now I want to translate the day rates. I wrote a function for this liske getDayRate(date,currency).
```
getDayRate <- function(date, currency) {
currencies[which(as.character(currencies[,c("Date")]) == date),c(currency)]
}

getDayRate("2013-06-20","EUR")
```
Now I want to apply getDayRate(date,currency) to each row of TRX so that for each row it uses the first and second element as arguments so I get teh ExchangeRate.

apply(x,1,fun()) does not work as it requires a matrix with numbers. In theory i would have to convert the dataframes to the indices and then use apply.

Is there a better way?
- dickoa almost 11 years
  
  Can you post the getDayRate function ?
- Thomas almost 11 years
  
  Take a look at mapply.
- dickoa almost 11 years
  
  +1 I think mapply is the way to go
user670186 almost 11 years

that should work, but I get a an error: Error in as.matrix.data.frame(X) : dims [product 5000] do not match the length of object [5009]
user670186 almost 11 years

problem is it seems apply converts the dataframe to a matrix. in that conversion the matrix is longer than my data frame and I have no idea why
Michele almost 11 years

try with adply from plyr. the actual problem is the everything in this question seems meaningless to me...
eddi almost 11 years

@user670186 I'm guessing that's because your function doesn't return a single number; I agree with Michele, it's unclear what you're trying to do - what exactly is ExchangeRate supposed to be? why do you have several dates per currency in your currencies table?