Why is SA2-by-DJZ-2011 re-reading?

[mattdowle]

In #2481, an aside to revisit was why the reread is happening. Row 0 is in-sample, so it shouldn't need a reread. Likely due to the (needed) header=FALSE.

fread("~/Downloads/SA2-by-DJZ-2011.csv", verbose=TRUE, header=FALSE)
...
  Type counts:
         1 : bool8     '1'
         1 : int32     '5'
         2 : string    'A'
=============================
   0.000s (  0%) Memory map 0.341GB file
   0.002s (  0%) sep=',' ncol=4 and header detection
   0.000s (  0%) Column type detection using 10027 sample rows
   0.196s ( 10%) Allocation of 25164895 rows x 4 cols (0.469GB) of which 22885380 ( 91%) rows used
   1.829s ( 90%) Reading 352 chunks of 0.993MB (64991 rows) using 8 threads
   =    0.000s (  0%) Finding first non-embedded \n after each jump
   +    0.358s ( 18%) Parse to row-major thread buffers (grown 0 times)
   +    0.713s ( 35%) Transpose
   +    0.757s ( 37%) Waiting
   0.592s ( 29%) Rereading 1 columns due to out-of-sample type exceptions
   2.027s        Total
Column 1 ("") bumped from 'bool8' to 'string' due to <<"Goulburn">> on row 0

[mattdowle]

No re-read is happening now, given recent improvements. Will add test and close.

[st-pasha]

Re-read is still happening:

> fread("~/Downloads/SA2-by-DJZ-2011.csv", verbose=T, header=F) -> f0
... (no re-read)

> fread("~/Downloads/SA2-by-DJZ-2011.csv", verbose=T, header=T) -> f0
...
Column 1 ("Goulburn") bumped from 'bool8' to 'string' due to <<"Bega-Eden Hinterland">> on row 1202

> fread("~/Downloads/SA2-by-DJZ-2011.csv", verbose=T, header=NA) -> f0
...
Column 1 ("") bumped from 'bool8' to 'string' due to <<"Goulburn">> on row 0