Improve the speed of displaying large datasets in Jupyter Notebook (#81)

fix #82 
* Improve the speed of displaying large datasets in Jupyter Notebook

Fix the issue that Jupyter Notebook takes too much time when displaying large datasets with many columns.

The getmaxwidths() function now only calculates the maximum width of the columns that will be displayed, not all columns.

Please see issue for more details.

* Modify the initialization code of maxwidths

Now might break the loop before filling all the values of maxwidths.

The initialize code of maxwidths has been modified.

Author: Kexi002

src/abstractdataset/io.jl

@@ -45,11 +45,13 @@ function getmaxwidths(ds::AbstractDataset,
                       show_eltype::Bool,
                       buffer::IOBuffer,
                       truncstring::Int)
-    maxwidths = Vector{Int}(undef, size(ds, 2) + 1)
+    maxwidths = zeros(Int, size(ds, 2) + 1)
 
     undefstrwidth = ourstrwidth(io, "#undef", buffer, truncstring)
 
     ct = show_eltype ? batch_compacttype(Any[eltype(c) for c in eachcol(ds)]) : String[]
+    tty_cols = displaysize(io)[2]
+    maxwidthsum = 0
     j = 1
     for (col_idx, (name, col)) in enumerate(pairs(eachcol(ds)))
         # (1) Consider length of column name
@@ -70,7 +72,12 @@ function getmaxwidths(ds::AbstractDataset,
         else
             maxwidths[j] = maxwidth
         end
+        maxwidthsum += (maxwidths[j] + 2)
         j += 1
+        # If the sum of column widths is already larger than COLUMNS, do not need calculate max width for the rest columns
+        if maxwidthsum >= tty_cols
+            break
+        end
     end
 
     # do not truncate rowlabel