Assigning NA to new column before a loop in R

Question

What do you think of assigning NA to new column before a loop? Is it consider a best practice ? Is there a more elegant way to do this?

I discovered don't assign NA value to columns before fill them with a loop can cause some troubles, specially on rows where API can't bring an answer: it fill the row with the data from the previous line...

Can you please help?

Url <- c("https://www.r-project.org/","https://cran.r-project.org/")
df <- data.frame(Url)

URL_row <- nrow(df)
df$PageSpeed_Score <- NA
df$PageSpeed_NumberResources <- NA
df$PageSpeed_NumberHosts <- NA

for (i in 1:URL_row) {

    url_to_check <- as.character(df[i, "Url"])
    print(url_to_check)

    PageSpeed_APIrequest <- paste("https://www.googleapis.com/pagespeedonline/v2/runPagespeed?url=", url_to_check,"&strategy=desktop", sep = "")
    PageSpeed_APIrequest <- fromJSON(PageSpeed_APIrequest)

    df$PageSpeed_Score[i] <- PageSpeed_APIrequest$rule$SPEED
    df$PageSpeed_NumberResources[i] <- PageSpeed_APIrequest$pageStats$numberResources
    df$PageSpeed_NumberHosts[i] <- PageSpeed_APIrequest$pageStats$numberHosts

}

Answer 1

The only (small) issue with NA is that it is of type logical.

typeof(NA)

It might be better to use NA_character_ or NA_real_ (etc.) depending on the expected type of the column. Not a big deal because mutability in R, though.