library(readr)
library(caret)
library(plyr)
library(dplyr)

cat("reading one record from the training data \n\n\n")
one_training_data<-read.csv('../input/train.csv', header=T, nrows=1)
cat("the 11 variable names in the training data \n\n\n")
names(one_training_data)


classes <- sapply(one_training_data, class)
cat("\n\n\n sample records of the training data \n\n\n\n")
head(one_training_data)

cat("\n\n\n reading thousand records from the training data \n\n\n")
big_training_data<-read.csv('../input/train.csv', header=T,colClasses = classes,nrows=100000)
cat("\n\n\n displaying using str \n\n\n")
str(big_training_data)
class(big_training_data)
dim(big_training_data)
filter(big_training_data,Semana=="3")
dim(filter)
cat("\n\n\n negative values and zero values\n\n\n")

cat("\n")
#one_testing_data<-read.csv('../input/test.csv', header=T, nrows=1)
#head(one_testing_data)
#classes1 <- sapply(one_testing_data, class)
#big_testing_data<-read.csv('../input/test.csv', header=T,colClasses = classes1,nrows=1000)
#str(big_testing_data)

#cat("\n\n using the caret library and the train function")
#model <- train(Demanda_uni_equil ~ Semana + Agencia_ID+Canal_ID+Ruta_SAK+Cliente_ID+Producto_ID,method="gbm",data=big_training_data, verbose=FALSE)
#print(model)


#cat("\n\n using the prdict function on the model already built\n\n\n")
#predicting <- predict(model, newdata = big_testing_data)
#str(predicting)

#submit <- data.frame(ID=big_testing_data$id, Demanda_uni_equil=predicting)
#names(submit)
#which(submit$Demanda_uni_equil<0)
#submit$Demanda_uni_equil[submit$Demanda_uni_equil < 0] <- 1
#write.csv(submit, "submit.csv", row.names = F)