Function readfile() doesn't read large files

[waqu]

I use this code to copy files:

Code: Select all

 writefile("I:\Temp\file.avi",readfile("D:\Temp\file.avi","b"),,"b");

Files size over 100 MB copied incompletely, but only the first 100Mb.

[highend]

Iirc, the limit is by design.

Why are you trying to copy an .avi file by readfile / writefile?

[waqu]

Iirc, the limit is by design.

Why are you trying to copy an .avi file by readfile / writefile?

I need to merge multiple files into one.

[highend]

Use the dos copy command instead (with run / runret):

Code: Select all

copy /b file1+file2+file3 targetFile

[PeterH]

You *could* also test for file size, and if too big, copy parts of the file in a loop.
(readfile() knows operands for numbytes and start)

But in general highends solution seems more straightforward

[highend]

It's a kind of operation I wouldn't do with XY just because you block it while the script is running and assembling large video files is (on non ssds) a time consuming process so it's better to externalize it (at least if you don't want to use a XY instance only for this task).

[waqu]

Use the dos copy command instead (with run / runret):

Code: Select all

copy /b file1+file2+file3 targetFile

I need after the merger split file of the same size.

[highend]

Then use PeterH's method or a command line splitting tool?

E.g.:
http://www.lawrencegoetz.com/programs/gfs101.zip

or
split from:
http://sourceforge.net/projects/unxutils/

[admin]

FYI, upcoming change log has a tip:

Code: Select all

    > SC readfile: This command will not read more than 100 MB at a time. To 
      read larger files use the "start" parameter and do it portion by portion.
      This information has been missing from the help so far. Will be added soon.

[SkyFrontier]

'start' method is obsolete/innacurate/inadequate since XY started distinguishing double-byte chars, isn't it?
(written completely by bad memory, as I stumbled on this often in the past)

[admin]

Not at all! In the case of double-byte chars "start" points to the character position. Easy and intuitive to use.

[SkyFrontier]

I recall I had problems because XY was auto-detecting and forcing readfile() to interpret sections as single or double chars, leading to said innacuracy... Can this be user driven?

[admin]

Did some tests: no problems visible.

[PeterH]

In the case of double-byte chars "start" points to the character position. Easy and intuitive to use.

Code: Select all

    > SC readfile: This command will not read more than 100 MB at a time. To
      read larger files use the "start" parameter and do it portion by portion.
      This information has been missing from the help so far. Will be added soon.
      Note that the "start" parameter points to the *character* position in case
      of text files (in double-byte encodings this is different from the byte
      position).

Docu (maybe old?) says:
numbytes: number of bytes to read
start: byte position to read from

OK: if numbytes and start are consistent, this "charcount" may be fine.

But note: if max 100MB may be read, numbytes (being chars!) *must* be sufficient smaller for multibyte charsets?
(Else it would be 100Mchars )

[admin]

Yes, Help needs and upgrade here...