Selecting columns in a DataFrame

Often it is useful to be able to select only a select number of columns in a dataframe, in conjunction or separately from filtering the rows of a dataframe. We can do this to simply re-size the dataframe to only what we need for a particular future analysis, or to isolate the part of a dataframe we need to perform other operations on.

Working With Data

Selecting Columns

To do this, we can select a particular set of columns. Let’s use the ice dataframe example from the lterdatasampler package:

# load packages 
library(lterdatasampler)
library(tidyverse)

# load in dataframe we'll work with
ice <- lterdatasampler::ntl_icecover

We can look at the columns in this dataframe:

names(ice)
## [1] "lakeid"       "ice_on"       "ice_off"      "ice_duration"
## [5] "year"

We may want to make a smaller dataframe with only the columns year and ice_duration . We can do this with the dplyr function select() :

small_ice <- ice %>%
dplyr::select(year, ice_duration)
names(small_ice)
## [1] "year"         "ice_duration"

And we can see this works as expected.

Ice Cover

When we think of climate change, we’re often treated to news media of melting glaciers or ice caps at the poles. However, reduction of ice cover is not unique to those environments, and in fact, many lake ecosystems rely on ice cover during the winter. Three lakes near Madison, WI, USA have been used for the monitoring of how climate change is possibly causing the change of ice cover duration in temperate lakes, and these data are the ones we use in this example.

Selecting & Filtering

It is always useful to be able to perform multiple operations on a dataframe at once, since it’s rarely the case that we only need to do a single thing before our data object is ready for analysis.

We can “pipe” our operations together using the dplyr pipes. For example, we could select the same columns as we did already, but also filter ice_duration at the same time. Let’s see what values that variable can take on:

# hist() will just give us a histogram of the values
hist(ice$ice_duration)

So we might want to filter our data to only values below 100. We can do that like this:

filtered_ice <- ice %>%
dplyr::filter(ice_duration < 100) %>%
dplyr::select(year, ice_duration)

Now we can look at the columns we have in this new object and also what values ice_duration takes on to check it worked correctly:

names(filtered_ice)
## [1] "year"         "ice_duration"
hist(filtered_ice$ice_duration)

We can see this worked as expected!