meta file for lab experiments

[nirgrahamuk]

library(magrittr)
tjt <- temp.jimported.tidy

for(y in  2:length(tjt)) {
     # start the unnest in col 2
    tjt %<>% unnest_longer(col = colnames(tjt[y]), 
                            indices_include = FALSE)}

?

[lsilva]

@nirgrahamuk

tjt <- temp.jimported.tidy

for(y in  2:length(tjt)) {
     # start the unnest in col 2
    tjt %<>% unnest_longer(col = colnames(tjt[y]), 
                            indices_include = FALSE)}

its cleaner -- I will use it!

But honestly, I was thinking of a lapply() style, where it would do the unnest_longer() for all the cols in a given data.frame where it would piped it self without a for loop; if it makes any sense.

[nirgrahamuk]

something to play with...

library(tidyverse)

intext <-'{  
   "Experiment1_01":[
     {
       "var1" : ["[[A01]]"],
       "var2" : ["2","1","3"],
       "var3" : ["0.2", "0.5"]
     }]
     ,
  "Experiment2_01":[
    {
      "var1" : ["[[A03]]"],
      "var2" : ["1"],
      "var3" : ["0.3","0.5"]
    }
    ]
}'
myt <- tempfile()
writeLines(intext,myt)
jimported <- jsonlite::read_json(myt, 
                                 simplifyVector = TRUE)  


jimported.df <- tibble()
if (length(jimported) > 1) { 
  for (k in 1:length(jimported)) {
    jimported.df <- bind_rows(jimported.df,
                                     cbind(tibble(exp=names(jimported[k])),jimported[[k]]))
    
  }
} else {
  jimported.df <-bind_rows(jimported.df,
                                  cbind(tibble(exp=names(jimported[1])),jimported[[1]])) 
  
}


#unnest all vars 
myvars<- names(jimported.df)[startsWith(names(jimported.df),"var")]
walk(myvars,
     ~assign("jimported.df",
             unnest_longer(data=jimported.df,
                           col = .x),
                           envir = .GlobalEnv)
     )

[HanOostdijk]

Nice @nirgrahamuk!
But I prefer your earlier solution. The latter is more to study than to play
But I also like the fact that you include the input: much easier for the reader

[lsilva]

So, I think it's all running for now.

Thanks for your help @HanOostdijk and @nirgrahamuk