13

I have a list of dataframes:

str(subsets.d)
List of 22
 $ 1       :'data.frame':   358 obs. of  118 variables:
  ..$ Ac_2017_1 : num [1:358] 0 0 0 0 0 0 0 0 0 0 ...
  ..$ Ac_9808_1 : num [1:358] 0 0 0 0 0 ...
  ..$ dates     : Ord.factor w/ 6 levels "April"<"May"<..: 1 1 1 1 1 1 1 
 $ 19      :'data.frame':   358 obs. of  2 variables:
  ..$ Ac_8598_19: num [1:358] 0.000257 0.000288 0.000171 0 0.000562 ...
  ..$ dates     : Ord.factor w/ 6 levels "April"<"May"<..: 1 1 1 1 1 1 1

I proceed by transforming this into long formats:

library(reshape2)
subsets.m <- lapply(subsets.d, function(x) melt(x))

and then i want to produce boxplots by dates:

library(ggplot2)
lapply(subsets.m, function(x)
  ggplot(x, aes(dates, value)) + geom_boxplot() + facet_wrap(~variable, scale="free_y"))

Is it possible to put a ggsave() or pdf() inside or otside the lapply? if so, how can i name the saved plots according to the dataframe?

Bonus question, since the dataframes inside the list have different dimensions, the plots can differ in size. Can i adjust for that?

Here is some small example data:

list1 <- list(
  data.frame(
    Ac_7595_JG37 = c(
      0.000128383, 0.000576914, 0.000341631, 0.000729133, 0.000187486, 0.00086127,
      0.000594978, 0.000631912, 0.000502274, 0.0004846, 0, 0.000148386, 0.000298153,
      0.000828969, 0.000436815, 7.28336e-05, 0.000304842, 0.000301909, 0.000233694,
      0.000208491
    ),
    Ac_7474_JG37 = c(
      0, 0, 0.000512446, 0, 0.000562457, 0, 0.000198326, 0.000473934, 0,
      0.000138457, 0.000132554, 0, 0.000198769, 0.000207242, 0.000145605,
      0.000364168, 0, 0.000301909, 0.000116847, 0
    ),
    dates = ordered(
      rep(c("April", "May"), each = 10L),
      levels = c("April", "May", "June", "August", "October", "November")
    )
  ),
  data.frame(
    Ac_8732_20 = c(
      0.000513534, 0.000384609, 0.000341631, 0.000729133, 0.000937429, 0.000322976,
      0.000297489, 0.000394945, 0.000669698, 0.000346143, 0.000265108, 0.000296773,
      0.000596306, 0.000310863, 0, 0.000509835, 0, 0.000301909, 0.000233694, 0
    ),
    dates = ordered(
      rep(c("April", "May"), each = 10L),
      levels = c("April", "May", "June", "August", "October", "November")
    )
  )
)
moodymudskipper
  • 46,417
  • 11
  • 121
  • 167
nouse
  • 3,315
  • 2
  • 29
  • 56
  • 2
    Use a `for` loop. It's the better choice for this task. – Roland Dec 12 '15 at 16:17
  • yes you can, and perhaps `deparse(substitute())` is an approach for naming. I have also used the approach of assigning a `comment` to each `data.frame` and using `comment(df)` in the save name. – MichaelChirico Dec 12 '15 at 16:19
  • 1
    What are the names in your example? And in what way do you want to "adjust" for the different sizes? – lukeA Dec 12 '15 at 16:22

1 Answers1

11

Your plots are stored in a list, hence you can use lapply on the output to save all of them. This has been answered here: Saving a list of plots by their names()

# create data for this example (data above too involved)
df <- data.frame(value = rnorm(100), dates = rep(1:50, 2), type = rep(c("a", "b")))
list1 <- split(df, df$type)


plots <- lapply(list1, function(x) ggplot(x, aes(dates, value)) + geom_boxplot())

lapply(names(plots), 
       function(x) ggsave(filename=paste(x,".jpeg",sep=""), plot=plots[[x]]))
Community
  • 1
  • 1
Raad
  • 2,675
  • 1
  • 13
  • 26