A histogram with the y-axis on the log scale will be a rather odd histogram. Technically it will still fit the definition, but it could look rather misleading: the peaks will be flattened relative to the rest of the distribution.
Instead of using a log transformation, have you considered:
Dividing the counts by 1 million:
h <- hist(hplot$V1, plot=FALSE)
h$counts <- h$counts/1e6
plot(h)
Plotting the histogram as a density estimate:
hist(hplot$V1, freq=FALSE)
