Filtering data in a dataframe based on criteria

问题

I am new to R and can't get to grips with this concept. Suppose I have a table loaded called "places" with 3 say columns - city, population and average summer temperature

Say I want to "filter" - produce a new table object where population is less than 1 million and average summer temperature is greater than 70 degrees.

In any other program I have used this would be pretty easy but having done some research I'm working myself up into greater confusion. Given the purpose of R and what it does this must be pretty simple stuff.

How would I apply the above conditions to the table? What would the steps be? From what i understand, I cannot easily just select the table headings based on their name, which would be nice (e.g. WHERE city < 1,000,000 )

回答1:

Given a dataframe "dfrm" with the names of the cities in the 'city' column, the population in the "population" column and the average summer temperature in the "meanSummerT" column your request for the subset meeting those joint requirements would be met with any of these:

subset( dfrm, population < 1e6 & meanSummerT > 70)
dfrm[ which(dfrm$population < 1e6 & dfrm$meanSummerT > 70) , ]
dfrm[ which(dfrm[['population']] < 1e6 & dfrm[['meanSummerT']] > 70) , ]

If you wanted just the names of the cities meeting those joint criteria then these would work:

subset( dfrm, population < 1e6 & meanSummerT > 70 , city)
dfrm[ which(dfrm$population < 1e6 & dfrm$meanSummerT > 70) , "city" ]
dfrm[ which(dfrm[['population']] < 1e6 & dfrm[['meanSummerT']] > 70) , "city" ]

Note that the column names are not quoted in the subset or following the "$" operator but they are quoted inside "[["

回答2:

You are looking for subset

if your data is called mydata

newdata <- subset(mydata, city < 1e6)

Or you could use [, which is programatically safer

newdata <- mydata[mydata$city < 1e6]

For more than one condition use & or | where approriate

You could also use the sqldf package to use sql

library(sqldf)

newdata <-  sqldf('select * from mydata where city > 1e6')

Or you could use data.table which makes the syntax easier for [ (as well as being memory efficient)

library(data.table)

mydatatable <- data.table(mydata)
newdata <- mydatatable[city > 1e6]

来源：https://stackoverflow.com/questions/14205583/filtering-data-in-a-dataframe-based-on-criteria

标签

subset