How to vectorize a for loop in R - r

I'm trying to clean this code up and was wondering if anybody has any suggestions on how to run this in R without a loop. I have a dataset called data with 100 variables and 200,000 observations. What I want to do is essentially expand the dataset by multiplying each observation by a specific scalar and then combine the data together. In the end, I need a data set with 800,000 observations (I have four categories to create) and 101 variables. Here's a loop that I wrote that does this, but it is very inefficient and I'd like something quicker and more efficient.
datanew <- c()
for (i in 1:51){
for (k in 1:6){
for (m in 1:4){
sub <- subset(data,data$var1==i & data$var2==k)
sub[,4:(ncol(sub)-1)] <- filingstat0711[i,k,m]*sub[,4:(ncol(sub)-1)]
sub$newvar <- m
datanew <- rbind(datanew,sub)
}
}
}
Please let me know what you think and thanks for the help.
Below is some sample data with 2K observations instead of 200K
# SAMPLE DATA
#------------------------------------------------#
mydf <- as.data.frame(matrix(rnorm(100 * 20e2), ncol=20e2, nrow=100))
var1 <- c(sapply(seq(41), function(x) sample(1:51)))[1:20e2]
var2 <- c(sapply(seq(2 + 20e2/6), function(x) sample(1:6)))[1:20e2]
#----------------------------------#
mydf <- cbind(var1, var2, round(mydf[3:100]*2.5, 2))
filingstat0711 <- array(round(rnorm(51*6*4)*1.5 + abs(rnorm(2)*10)), dim=c(51,6,4))
#------------------------------------------------#

You can try the following. Notice that we replaced the first two for loops with a call to mapply and the third for loop with a call to lapply.
Also, we are creating two vectors that we will combine for vectorized multiplication.
# create a table of the i-k index combinations using `expand.grid`
ixk <- expand.grid(i=1:51, k=1:6)
# Take a look at what expand.grid does
head(ixk, 60)
# create two vectors for multiplying against our dataframe subset
multpVec <- c(rep(c(0, 1), times=c(4, ncol(mydf)-4-1)), 0)
invVec <- !multpVec
# example of how we will use the vectors
(multpVec * filingstat0711[1, 2, 1] + invVec)
# Instead of for loops, we can use mapply.
newdf <-
mapply(function(i, k)
# The function that you are `mapply`ing is:
# rbingd'ing a list of dataframes, which were subsetted by matching var1 & var2
# and then multiplying by a value in filingstat
do.call(rbind,
# iterating over m
lapply(1:4, function(m)
# the cbind is for adding the newvar=m, at the end of the subtable
cbind(
# we transpose twice: first the subset to multiply our vector.
# Then the result, to get back our orignal form
t( t(subset(mydf, var1==i & mydf$var2==k)) *
(multpVec * filingstat0711[i,k,m] + invVec)),
# this is an argument to cbind
"newvar"=m)
)),
# the two lists you are passing as arguments are the columns of the expanded grid
ixk$i, ixk$k, SIMPLIFY=FALSE
)
# flatten the data frame
newdf <- do.call(rbind, newdf)
Two points to note:
Try not to use words like data, table, df, sub etc which are commonly used functions
In the above code I used mydf in place of data.
You can use apply(ixk, 1, fu..) instead of the mapply that I used, but I think mapply makes for cleaner code in this situation

Related

Compute 15 rows in parallel (through vectorization) and create df with them

I am creating 15 rows in a dataframe, like this. I cannot show my real code, but the create row function involves complex calculations that can be put in a function. Any ideas on how I can do this using lapply, apply, etc. to create all 15 in parallel and then concatenate all the rows into a dataframe? I think using lapply will work (i.e. put all rows in a list, then unlist and concatenate, but not exactly sure how to do it).
for( i in 1:15 ) {
row <- create_row()
# row is essentially a dataframe with 1 row
rbind(my_df,row)
}
Something like this should work for you,
create_row <- function(){
rnorm(10, 0,1)
}
my_list <- vector(100, mode = "list")
my_list_2 <- lapply(my_list, function(x) create_row())
data.frame(t(sapply(my_list_2,c)))
The create_row function is just make the example reproducible, then we predefine an empty list, then fill it with the result from the create_row() function, then convert the resulting list to a data frame.
Alternatively, predefine a matrix and use the apply functions, over the row margin, then use the t (transpose) function, to get the output correct,
df <- data.frame(matrix(ncol = 10, nrow = 100))
t(apply(df, 1, function(x) create_row(x)))

R: Performance issue when finding maximum of splitted list

When trying to find the maximum values of a splitted list, I run into serious performance issues.
Is there a way I can optimize the following code:
# Generate data for this MWE
x <- matrix(runif(900 * 9000), nrow = 900, ncol = 9000)
y <- rep(1:100, each = 9)
my_data <- cbind(y, x)
my_data <- data.frame(my_data)
# This is the critical part I would like to optimize
my_data_split <- split(my_data, y)
max_values <- lapply(my_data_split, function(x) x[which.max(x[ , 50]), ])
I want to get the rows where a given column hits its maximum for a given group (it should be easier to understand from the code).
I know that splitting into a list is probably the reason for the slow performance, but I don't know how to circumvent it.
This may not be immediately clear to you.
There is an internal function max.col doing something similar, except that it finds position index of the maximum along a matrix row (not column). So if you transpose your original matrix x, you will be able to use this function.
Complexity steps in when you want to do max.col by group. The split-lapply convention is needed. But, if after the transpose, we convert the matrix to a data frame, we can do split.default. (Note it is not split or split.data.frame. Here the data frame is treated as a list (vector), so the split happens among the data frame columns.) Finally, we do an sapply to apply max.col by group and cbind the result into a matrix.
tx <- data.frame(t(x))
tx.group <- split.default(tx, y) ## note the `split.default`, not `split`
pos <- sapply(tx.group, max.col)
The resulting pos is something like a look-up table. It has 9000 rows and 100 columns (groups). The pos[i, j] gives the index you want for the i-th column (of your original non-transposed matrix) and j-th group. So your final extraction for the 50-th column and all groups is
max_values <- Map("[[", tx.group, pos[50, ])
You just generate the look-up table once, and make arbitrary extraction at any time.
Disadvantage of this method:
After the split, data in each group are stored in a data frame rather than a matrix. That is, for example, tx.group[[1]] is a 9000 x 9 data frame. But max.col expects a matrix so it will convert this data frame into a matrix internally.
Thus, the major performance / memory overhead includes:
initial matrix transposition;
matrix to data frame conversion;
data frame to matrix conversion (per group).
I am not sure whether we eliminate all above with some functions from MatrixStats package. I look forward to seeing a solution with that.
But anyway, this answer is already much faster than what OP originally does.
A solution using {dplyr}:
# Generate data for this MWE
x <- matrix(runif(900 * 9000), nrow = 900, ncol = 9000)
y <- rep(1:100, each = 9)
my_data <- cbind.data.frame(y, x)
# This is the critical part I would like to optimize
system.time({
my_data_split <- split(my_data, y)
max_values <- lapply(my_data_split, function(x) x[which.max(x[ , 50]), ])
})
# Using {dplyr} is 9 times faster, but you get results in a slightly different format
library(dplyr)
system.time({
max_values2 <- my_data %>%
group_by(y) %>%
do(max_values = .[which.max(.[[50]]), ])
})
all.equal(max_values[[1]], max_values2$max_values[[1]], check.attributes = FALSE)

R: Use apply correctly

I would like to create a simple code to join columns then count how many category variable has in my data frame. My problems is when I use apply, I get a right result but five times.
EXAMPLE:
a <- c('car','bike',NA,'moto','skate')
b <- c(NA,'car',NA,NA,'bike')
c <- c('car',NA,NA,'skate',NA)
d <- c('moto','skate',NA,'car',NA)
data <- data.frame(a,b,c,d)
then, using apply:
x <- vector('list',length = NCOL(data)*NROW(data))
one_column <- apply(data,1,function(y){
x <- rbind(y,x)
return(x)
})
Then unlist and use table for count how many categorical variables I have in my data:
one_column <- unlist(one_column)
table(one_column)
But for I get the right result I need divide by 5:
table(one_column)/5
The x vector you created is 5 times larger than each column, so it's recycling the data. What you want is this instead.
x <- vector('list',length = NCOL(data))
Or, like emilliman5 says, just use table(unlist(data)).

Split data to make train and test sets - for loop - insert variable to subset by row

I am trying to subset this data frame by pre determined row numbers.
# Make dummy data frame
df <- data.frame(data=1:200)
train.length <- 1:2
# Set pre determined row numbers for subsetting
train.length.1 = 1:50
test.length.1 = 50:100
train.length.2 = 50:100
test.length.2 = 100:150
train.list <- list()
test.list <- list()
# Loop for subsetting by row, using row numbers in variables above
for (i in 1:length(train.length)) {
# subset by row number, each row number in variables train.length.1,2etc..
train.list[[i]] <- df[train.length.[i],] # need to place the variable train.length.n here...
test.list[[i]] <- df[test.length.[i],] # place test.length.n variable here..
# save outcome to lists
}
My question is, if I have my row numbers stored in a variable, how I do place each [ith] one inside the subsetting code?
I have tried:
df[train.length.[i],]
also
df[paste0"train.length.",[i],]
however that pastes as a character and it doesnt read my train.length.n variable... as below
> train.list[[i]] <- df[c(paste0("train.length.",train.length[i])),]
> train.list
[[1]]
data data1
NA NA NA
If i have the variable in there by itself, it works as intended. Just need it to work in a for loop
Desired output - print those below
train.set.output.1 <- df[train.length.1,]
test.set.output.1 <- df[test.length.1,]
train.set.output.2 <- df[train.length.2,]
test.set.output.2 <- df[test.length.2,]
I can do this manually, but its cumersome for lots of train / test sets... hence for loop
Consider staggered seq() and pass the number sequences in lapply to slice by rows. Also, for equal-length dataframes, you likely intended starts at 1, 51, 101, ...
train_num_set <- seq(1, 200, by=50)
train.list <- lapply(train_num_set, function(i) df[c(i:(i+49)),])
test_num_set <- seq(51, 200, by=50)
test.list <- lapply(test_num_set, function(i) df[c(i:(i+49)),])
Create a function that splits your data frame into different chunks:
split_frame_by_chunks <- function(data_frame, chunk_size) {
n <- nrow(data_frame)
r <- rep(1:ceiling(n/chunk_size),each=chunk_size)[1:n]
sub_frames <- split(data_frame,r)
return(sub_frames)
}
Call your function using your data frame and chunk size. In your case, you are splitting your data frame into chunks of 50:
chunked_frames <- split_frame_by_chunks(data_frame, 50)
Decide number of train/test splits to create in the loop
num_splits <- 2
Create the appropriate train and test sets inside your loop. In this case, I am creating the 2 you showed in your question. (i.e. the first loop creates a train and test set with rows 1-50 and 50-100 respectively):
for(i in 1:num_splits) {
this_train <- chunked_frames[i]
this_test <- chunked_frames[i+1]
}
Just do whatever you need to the dynamically created train and test frames inside your loop.

Calculating log returns over columns of a data frame + store the results in a new data frame

My data frame contains 22 columns: "DATE", "INDEX" and S1, S2, S3 ... S20. There are over 4322 rows. I want to calculate log returns and store the results in a data frame. That should give me 4321 rows.
I run this code, but I am sure there is a much more elegant way to do the calculation in a short way.
# count the sum of rows in order to make the following formula work appropriately - (n-1)
n <- nrow(df)
# calculating the log returns (natural logarithm), of INDEX and S1-20
LogRet_INDEX <- log(df$INDEX[2:n])-log(df$INDEX[1:(n-1)])
LogRet_S1 <- log(df$S1[2:n])-log(df$S1[1:(n-1)])
LogRet_S2 <- log(df$S2[2:n])-log(df$S2[1:(n-1)])
LogRet_S3 <- log(df$S3[2:n])-log(df$S3[1:(n-1)])
LogRet_S4 <- log(df$S4[2:n])-log(df$S4[1:(n-1)])
LogRet_S5 <- log(df$S5[2:n])-log(df$S5[1:(n-1)])
LogRet_S6 <- log(df$S6[2:n])-log(df$S6[1:(n-1)])
LogRet_S7 <- log(df$S7[2:n])-log(df$S7[1:(n-1)])
LogRet_S8 <- log(df$S8[2:n])-log(df$S7[1:(n-1)])
LogRet_S9 <- log(df$S9[2:n])-log(df$S8[1:(n-1)])
LogRet_S10 <- log(df$S10[2:n])-log(df$S10[1:(n-1)])
LogRet_S11 <- log(df$S11[2:n])-log(df$S11[1:(n-1)])
LogRet_S12 <- log(df$S12[2:n])-log(df$S12[1:(n-1)])
LogRet_S13 <- log(df$S13[2:n])-log(df$S13[1:(n-1)])
LogRet_S14 <- log(df$S14[2:n])-log(df$S14[1:(n-1)])
LogRet_S15 <- log(df$S15[2:n])-log(df$S15[1:(n-1)])
LogRet_S16 <- log(df$S16[2:n])-log(df$S16[1:(n-1)])
LogRet_S17 <- log(df$S17[2:n])-log(df$S17[1:(n-1)])
LogRet_S18 <- log(df$S18[2:n])-log(df$S18[1:(n-1)])
LogRet_S19 <- log(df$S19[2:n])-log(df$S19[1:(n-1)])
LogRet_S20 <- log(df$S20[2:n])-log(df$S20[1:(n-1)])
# adding the results from the previous calculation (log returns) to a data frame
LogRet_df <- data.frame(LogRet_INDEX, LogRet_S1, LogRet_S2, LogRet_S3, LogRet_S4, LogRet_S5, LogRet_S6, LogRet_S7, LogRet_S8, LogRet_S9, LogRet_S10, LogRet_S11, LogRet_S12, LogRet_S13, LogRet_S14, LogRet_S15, LogRet_S16, LogRet_S17, LogRet_S18, LogRet_S19, LogRet_S20)
Is there a possibility to make this code shorter? Maybe some kind of loop or using a for argument? Since I am quite new to R, I try to improve my knowledge.
Any kind of help is highly appreciated!
You can use sapply to apply a function to each column of the data.frame.
What the code below does, is 1) take columns 2 to 22 from the data frame called df. 2) for each of this columns, calculate logarithm of the respective column and then calculate the difference between two neighboring rows. 3) when done, convert it to data.frame called df2
df2 <- as.data.frame(sapply(df[2:22], function(x) diff(log(x))))

Resources