@swasheck I think this guy has a better grasp of opinion v fact and the dangers of blurring the lines - while reading it I wondered if you'd agree

@AndriyM Gosh thanks.

The pleasure is all mine, Paul.

All the same. That was unexpected, very generous of you, and I appreciate it.

@ypercube Do you think this is real? politico.eu/article/varoufakis-plot-to-bring-back-drachma

Someone care to check my math(s)? A 500GiB file, is 536871 MB. Assuming a 5200 RPM drive at 50 MB/s, is it really going take almost 3 hours to read that (assuming no other disk contention, sequential reads, etc)

@JamesLupolt I seriously doubt it.

They probably had considered other plans but raiding banks? No.

@billinkc The unit conversions make sense but 50 MB/sec seems quite low for sequential throughput with a large block size, even for a single 5400 RPM drive.

@ypercube yeah I thought I would ask but it seems quite movie-plot and unlikely

@JamesLupolt the plot has been in Greek papers for a week or more.

I was running with the low-end range from overclockers.com/forums/showthread.php/… (second post) Was trying to give context as to why using XML to store 500GB of data was a "Bad Thing" (tm) stackoverflow.com/questions/31625660/…

I guess the international press is catching up ;)

@billinkc I think it's a valid point. You could get much worse (a few MB/sec) if it's purely random I/O with small block sizes on a small number of disks, though.

Is it common to get a 500 GB XML file as a data source to load?

I hope not

Presumably you have to read the whole thing to figure out whether it's well formed, before you can even do anything with it?

I wonder what's producing these giant XML files

You have two choices, you can basically read each row (as they're doing) and just assuming that it's good or maybe it reads until it closes the tag. The other is DOM parsing where it has to read the whole thing to make sense of it

I bet it's an export from a database and someone just said we'll use XML