CODEWITHSUNDEEP
rfor-loopdataframerbind
Sander Van der Zeeuw
Sander Van der Zeeuw

Reputation: 1092

Loop over data frame rows and create data frame from new rows

I'm trying to loop over the rows of a DF. While looping over the rows i want to do some changes to some rows, and create a new dataframe containing the new rows.

The dataframe i use looks like can be created with this:

df <- structure(
  list(
    campaign_name = c(
      "Category> fanshop",
      "Category> trainingspakken",
      "Category> trainingsshirts",
      "Category> hoodies",
      "Category> broeken",
      "Category> voetbalshirts"
    ),
    ad_group = c(
      "Pro team X[B]",
      "Pro team X[B]",
      "Pro team X[B]",
      "Pro team X[B]",
      "Pro team X[B]",
      "Pro team X[B]"
    ),
    category = c(
      "fanshop",
      "trainingspakken",
      "trainingsshirts",
      "hoodies",
      "broeken",
      "voetbalshirts"
    ),
    Final_URL = c(
      "https://fanshop/Pro-team-X.html",
      "https://fanshop/Pro-team-X/_Trainingspakken",
      "https://fanshop/Pro-team-X/_Trainingsshirts",
      "https://fanshop/Pro-team-X/_Hoodies_Sweaters",
      "https://fanshop/Pro-team-X/_Korte-broeken_Lange-broeken",
      "https://fanshop/Pro-team-X/_Voetbalshirts"
    ),
    team_name = c(
      "Pro team X",
      "Pro team X",
      "Pro team X",
      "Pro team X",
      "Pro team X",
      "Pro team X"
    ),
    keyword = c(
      "+Pro +team +X +fanshop",
      "+Pro +team +X +trainingspakken",
      "+Pro +team +X +trainingsshirts",
      "+Pro +team +X +hoodies",
      "+Pro +team +X +broeken",
      "+Pro +team +X +voetbalshirts"
    ),
    Criterion_type = c("Broad", "Broad", "Broad", "Broad", "Broad", "Broad")
  ),
  .Names = c(
    "campaign_name",
    "ad_group",
    "category",
    "Final_URL",
    "team_name",
    "keyword",
    "Criterion_type"
  ),
  row.names = c("1", "2", "3", "4", "5", "6"),
  class = "data.frame"
)

If i use the function below the rows print nicely and are changed. But as soon as i try to assign it to a data frame, off course it will be overwritten every time the loop runs.

for ( row in 1:nrow(df)) {
  temp_row <- df[row,]
  if (temp_row$Criterion_type == "Broad") {
    temp_row$keyword <- gsub("\\+", "", temp_row$keyword)
    temp_row$Criterion_type <- "Negative Exact"
  }
  print(temp_row)
}

After looking into many questions here and try many of the approaches i still did not manage to get it done properly. Much appreciated!

I expect each row to be modified according to the IF statement above. 1 row would look like this:

campaign_name   ad_group    category    Final_URL   team_name   keyword Criterion_type
Category> voetbalshirts Pro team X[B]   voetbalshirts   https://fanshop/Pro-team-X/_Voetbalshirts   Pro team X  paris saint germain voetbalshirts   Negative Exact

Some of the questions I tried already:

How to append rows to an R data frame

duplicate rows and create new data frame in R

Upvotes: 0

Views: 1862

Answers (1)

coffeinjunky
coffeinjunky

Reputation: 11514

Try this: 

library(dplyr)
new_df <- data.frame()
for ( row in 1:nrow(df)) {
  temp_row <- df[row,]
  if (temp_row$Criterion_type == "Broad") {
    new_df <- bind_rows(new_df, data.frame(keyword=gsub("\\+", "", temp_row$keyword), Criterion_type = "Negative Exact"))
  }
}

This gives you a new dataframe as follows:

new_df

                              keyword Criterion_type
1         paris saint germain fanshop Negative Exact
2 paris saint germain trainingspakken Negative Exact
3 paris saint germain trainingsshirts Negative Exact
4         paris saint germain hoodies Negative Exact
5         paris saint germain broeken Negative Exact
6   paris saint germain voetbalshirts Negative Exact

As a remark, though, note that you can achieve this probably much easier and potentially faster (since vectorized). For instance, 

df$keyword <- with(df, 
                   ifelse(Criterion_type=="Broad", gsub("\\+", "", keyword), keyword))
df$Criterion_type <- with(df, 
                   ifelse(Criterion_type=="Broad", "Negative Exact", Criterion_type))

achieves the same and is much more readable.

Upvotes: 1

Related Questions