New post
Discussion (11)2
Log in or sign up to continue
Robert Barbiaux · Feb 19

Yes, ^SPOOL is the simplest way to achieve this. If you need a string rather than a global, you can just get all lines from ^SPOOL, for example :

ClassMethod ZWriteToString() As %String
{
 #Dim result as %String
 #Dim i,lineCount as %Integer
 
 kill ^SPOOL($j)
 open 2:$j
 use 2
 zwrite
 s result=""
 s lineCount=$select($data(var):$za-1,1:$za-2)
 close 2
 for i=1:1:lineCount s result=result_^SPOOL($j,i)	
 return result
}
2 0
Alexey Maslov · Feb 22

Yeah, but it should be used with precautions. Let's see what happens when two processes access spooling device concurrently: 

USER>f {q:'($zh\1#20)} k a o 2 f i=1:1:10 {s a(i)=i*100} u 2 zw a c 2 ; process #1
USER>s i="" f  {s i=$o(^SPOOL(1,i),1,line) q:i=""  w line} ; look inside ^SPOOL(1)...
a(1)=100
a(2)=200
a(3)=300
a(4)=400
a(5)=500
a(6)=600
a(7)=700
a(8)=800
a(9)=900
a(10)=1000
{66892,65205{11
USER>f {q:'($zh\1#20)} k a o 2 f i=1:1:10 {s a(i)=i} u 2 zw a c 2 ; process #2
USER>s i="" f  {s i=$o(^SPOOL(1,i),1,line) q:i=""  w line} ; look inside ^SPOOL(1)...
a(1)=100
a(2)=200
a(3)=300
a(4)=400
a(5)=500
a(6)=600
a(7)=700
a(8)=800
a(9)=900
a(10)=1000
{66892,65205{11

As you can notice, one process's output suppressed the other's one. To avoid it, firstly RTFM: OPEN and USE Commands for Spooling Device, and implement some synchronization pattern, e.g. 

USER>f {q:'($zh\1#20)} s docIdx=$i(^SPOOL) k a o 2:(docIdx) f i=1:1:10 {s a(i)=i*100} u 2 zw a c 2 ; process #1
USER>s i="" f  {s i=$o(^SPOOL(docIdx,i),1,line) q:i=""  w line}
a(1)=100
a(2)=200
a(3)=300
a(4)=400
a(5)=500
a(6)=600
a(7)=700
a(8)=800
a(9)=900
a(10)=1000
{66892,66645{11{
USER>f {q:'($zh\1#20)} s docIdx=$i(^SPOOL) k a o 2:(docIdx) f i=1:1:10 {s a(i)=i} u 2 zw a c 2 ; process #2
USER>s i="" f  {s i=$o(^SPOOL(docIdx,i),1,line) q:i=""  w line}
a(1)=1
a(2)=2
a(3)=3
a(4)=4
a(5)=5
a(6)=6
a(7)=7
a(8)=8
a(9)=9
a(10)=10
{66892,66645{11{

Each process has got it's own output in ^SPOOL(docIdx) now. This approach works if every consumer of spooling facility follows the same pattern; incrementing ^SPOOL is just an easiest approach, while it would be better avoid touching system globals and increment something else according to your taste.

1 0
Alexey Maslov · Feb 23

^SPOOL(docIdx) approach allows spooling sharing across ECP network, while ^SPOOL($j) does not :)

%IS is an utility which allows to choose devices by CHUI routines users and %SPOOL utility can manipulate with spool files opened using the %IS utility only.

It seems that initial purpose of spooling facility has gone with such a users and such a printers (strictly character based, w/o ability of font selection, etc), while its usage still can be actual in rather exotic cases like this one.

0 0
Mike.W · Feb 26

Hi,

After many years of development and support I've become wary of one-line requests. 🤔 I wonder why you need this (Five whys - Wikipedia). For example, if you are trying to debug a mysterious state in a background job then maybe you just need "D LOG^%ETN" to store the variables in the error log. Or, at least you could look in there for ways to use $ORDER and $QUERY  to scan local variables without involving ^SPOOL (or, as we once did, opening a file to ZW to, closing, and then reading it back).

0 0