Question

I have my list MF wich contains 105 lists. Every list, MF[[1]] MF[[2]] …. MF[[105]] contains different number of data frames. Thus, MF[[1]][[1]] exists but MF[[1]][[2]] is NULL because there is just one data frame for MF[[1]]. In the other hand MF[[2]] contains 15 different data frames, so MF[[2]][[1]] to MF[[2]][[15]] exists.

The colnames of all the data frames in every 105 list is:

[1] "Run"             "Fecha"           "Serie"           "Patrimonio"      "Ret Log Pat"     "Percentil 5%"    "Percentil Monto"

I’m gonna ask my question with a concrete example. Let’s use MF[[2]] wich contains 15 different data frames. Here are some headers of those data frames:

head(MF[[2]][[1]]):
 Run      Fecha Serie Patrimonio  Ret Log Pat Percentil 5% Percentil Monto
31 8011 2002-08-18     1 4191689227 -0.456258862   -0.1973659      1305605031
32 8011 2002-08-19     1 4749171865  0.124866449   -0.2179453       913558775
33 8011 2002-08-20     1 5132656241  0.077653052   -0.2179453      1035059470
34 8011 2002-08-21     1 5088469783 -0.008646158   -0.2179453      1118638070
35 8011 2002-08-22     1 4998945148 -0.017750234   -0.2179453      1109007841
36 8011 2002-08-23     1 5449454077  0.086288515   -0.2179453      1089496372

head(MF[[2]][[2]])
    Run      Fecha Serie Patrimonio   Ret Log Pat Percentil 5% Percentil Monto
31 8011 2006-05-09   100 6413583009 -0.0076314490  -0.07046562       455399234
32 8011 2006-05-10   100 6412446421 -0.0001772315  -0.07046562       451937105
33 8011 2006-05-11   100 6380254435 -0.0050328784  -0.07046562       451857014
34 8011 2006-05-12   100 6381112038  0.0001344061  -0.07046562       449588586
35 8011 2006-05-13   100 6381970402  0.0001345073  -0.07046562       449649018
36 8011 2006-05-14   100 6315827940 -0.0104180360  -0.07046562       449709503

head(MF[[2]][[3]])
    Run      Fecha Serie Patrimonio   Ret Log Pat Percentil 5% Percentil Monto
31 8011 2002-08-18     2 3147993667 -0.0395416467  -0.03216529       105340167
32 8011 2002-08-19     2 3065335420 -0.0266083198  -0.03778848       118957901
33 8011 2002-08-20     2 3044946268 -0.0066737439  -0.03778848       115834372
34 8011 2002-08-21     2 3089802537  0.0146239300  -0.03778848       115063897
35 8011 2002-08-22     2 3090714960  0.0002952578  -0.03778848       116758947
36 8011 2002-08-23     2 3230667973  0.0442864759  -0.03778848       116793426

What I want is an iteration or whatever, that matches the column "Fecha" ( which means "Date" by the way), and if the Date matches, calculates the percent which represent each row of the column "Patrimonio" over the total sum of "Patrimonio" in which date matches.

Example given:

In this case we got:

head(MF[[2]][[1]]):
     Run      Fecha Serie Patrimonio  Ret Log Pat Percentil 5% Percentil Monto
    31 8011 2002-08-18     1 4191689227 -0.456258862   -0.1973659      1305605031

 head(MF[[2]][[3]])
        Run      Fecha Serie Patrimonio   Ret Log Pat Percentil 5% Percentil Monto
    31 8011 2002-08-18     2 3147993667 -0.0395416467  -0.03216529       105340167

So, MF[[2]][[1]][1,2]==MF[[2]][[3]][1,2] ( Dates matches ), then I want a new column over each data frame like this:

  New column for MF[[2]][[1]] = MF[[2]][[1]][1,4]/(MF[[2]][[1]][1,4]+MF[[2]][[3]][1,4]) =  4191689227/( 4191689227+ 3147993667) ( Percent Calculation over "Patrimonio" column )

 New column for MF[[2]][[3]] = MF[[2]][[3]][1,4]/(MF[[2]][[1]][1,4]+MF[[2]][[3]][1,4]) =  3147993667/( 4191689227+ 3147993667) ( Percent Calculation over "Patrimonio" column )

The thing is that I must match all the 15 data frames to calculate the "Patrimonio" percent by the variable "Fecha" and so on for all the 105 lists. Hope my doubt is clear enough.

Answer 1

I can’t easily use your data due to the “5%” in the headers. However, you need to use the apply family for the first step.

lapply(MF, yourfunction)

will apply yourfunction to each element of ML. Since each element of ML is a list also, you could lapply again (either in yourfunction or lapply(MF, lapply, yourfunction).

yourfunction will be something that works to do the calculation you want on a single data.frame. I find it easiest to extract one from these nested structures and write a function that works for it. Then worry about applying it to all the members of nested lists.

It sortof sounds like you’re wanting to compare the dates between data.frames. IF this is the case, your best bet is to combine them into a single frame rather than nested in a list.

You can do this in a few ways, but I like plyr.

library(plyr)
ldply(unlist(MF, recursive=FALSE), as.data.frame)

Then the comparisons are much more straight forward.