Conditional replacement of values in a data.frame

[Gouri Shankar Mishra]

Hi All

I want to conditional replace the values of a vector (df$var2) based on values of df$var1. Usually, I would write that as

df$var2 [df$var1 == "x"] <- df$var3 [df$var1 == "x"]

However, the above code looks at the value of var1 in the same row. I want to look either at a latter row or a prior row. Which row is again conditional. Which functions should I look at?

Specifically, in the attached excel, where var2 gives the travel purpose of a travel, and var1 identifies a location as an "anchor" location or not.

• Travel1 (Activity #4) is a home to office trip, hence travel purpose (var2) is based on activity in office (var3, "work").
  
• Similarly Travel2 (Activity#6) is a travel to work.
• In contrast, Travel3 (Activity#10) depends upon prior value of var2 because the destination location is Home. Hence the travel purpose is based on activity in the prior anchor location (work in office).
  

[Noam Ross]

Tricky! To clarify, I think the problem is the following:

TravelPurpose should be “Work” IF

• The LAST location which is an anchor location is “Office” OR
• The Previous location which is an anchor location is “Office”

This seems like a good place to break out the old for loop, as each iteration depends on values before and after it. There may be a more elegant vectorized solution, but this works:

# Load data and empty var1. Make sure to encode strings as characters, factors
# will mess up the logic below
df = read.csv("DRUGJul2.csv", stringsAsFactors=FALSE)
df$var1=character(11)                                 

# Create vectors with the indices of anchor locations and travel activities
anchors = which(df$var2==1) 
travels = which(df$var3=="Travel") 

#Iterate through "travel" locations
for(i in travels) {
  last_anchor = df$location[max(anchors[anchors < i])]  #Get last anchor loc
  next_anchor = df$location[min(anchors[anchors > i])]  #Get next anchor loc
  if("Office" %in% c(last_anchor, next_anchor)) {       #Are either "Office"?
    df$var1[i] = "Work"
  }
}

You’ll get warnings if you have “Travel” values in the first and last rows but it should still work.

--
Check out our R resources at http://www.noamross.net/davis-r-users-group.html
---
You received this message because you are subscribed to the Google Groups "Davis R Users' Group" group.
To unsubscribe from this group and stop receiving emails from it, send an email to davis-rug+...@googlegroups.com.
Visit this group at http://groups.google.com/group/davis-rug.
For more options, visit https://groups.google.com/d/optout.

​

[Gouri Shankar Mishra]

Thanks Noam