AltME groups: search

Anton, ah, but my function doesn't work with other action! values 
anyway:

>> a?: func [:value] [action! = type? value]
>> a? first

** Script Error: value expected series argument of type: series pair 
event money date objec
t port time tuple any-function library struct ...

Trying dobble get-word! :

>> a? :first
== false

So maybe this doesn't work with any action! datatype?

Is my understanding of find/part wrong, or is this a bug?
     series: ["a" "b" "c" "d" "e"]
     find/part series "c" at series 3  

     == none      ;; fails to match using the [at series 3] part of  'series
    print mold at series 3  
    == ["c" "d" "e"]  ;; but [at series3] does contain the match

The /part <range> argument (at series 3) is supposed to mark the 
end of the search range. So, here FIND is searching in ["a" "b"] 
only.

>> find/part series "c" 3
== ["c" "d" "e"]
>> find/part series "c" at series 3
== none
>> index? at series 3
== 3

This looks like an inconsistency to me.

Thanks. What was confusing me is an inconsistency between R2 and 
R3 -- was not sure if R3 was the bug fix or the problem.
R2:
    select/part series "c" at series 4
    == "d"
R3: 
    >> select/part series "c" at series 4
     == none

*** Do we agree R3 is right in this case?

In R3:
>> help swap
USAGE:
        SWAP series1 series2

DESCRIPTION:
        Swaps elements of a series. (Modifies)
        SWAP is an action value.

ARGUMENTS:
        series1 (series! gob!)
        series2 (series! gob!)
>> swap a: "abc" skip a 2 a
== "cba"

Also, the help text for as-string says "Coerces any type of string 
...". Should that read "any type of series"?

this value is only the basis for an intense series of string manipulations 
which make even the original data useless unless someone has the 
exact algorythm which generated the key.

From the script header: Purpose: "Working around string series"...I 
think that this is shooting yourself in the foot.

IIRC, Carl explained (at least) once that this behaviour is different 
on purpose. Using path notation with an index value or PICK should 
return NONE while using the prefix notation FIRST, SECOND,...LAST 
should return an error. So you have the choice to either silently 
handle missing values in series or raise an error!.

found a bug in 'FOR when used in series... it becomes an endless 
loop..

It is a bit confusing, also because both path! and file! are series, 
so it's easy to think, they should work the same.

>> series? d
== true
>> series? %/a/b/c
== true

A question: can it be justified, that file! is a series?

>> a: %/c
== %/c
>> first a
== #"/"
>> a/1
== %/c/1


Is file! a pseudo-series? What exactly defines a series? It's some 
sort of a sequence of values, that can be picked individually. And 
series can be sorted, right? So the methods (or functions), that 
can be performed on the datatype defines its type. What defines a 
series in REBOL?

each series type handles path actions how it wants to.

files append to themselves, blocks do lookup, strings return chars.


to be a series, in has to support insert, copy, first  which is why 
tuple isn't a serie.

Geomol: "What defines a series in REBOL?"


A series has contents, sequence and persistent position. The characters 
that make up a file (contents) are in a specific order (sequence) 
and you can do an offset reference to a later offset in the filename 
(position). If you do FIRST %a it will return #"a" every time (persistent 
position). Being a series doesn't mean that it is sortable, though 
many are. Being sortable means also having the contents be comparable, 
and the series be modifiable.


We were careful to make sure that things that aren't really series-like 
were removed from the series! typeset in R3. R2 is less consistent 
in this regard, so you have some types attempting to act series-like, 
poorly - pots being the worst example.


Some of the functions that act on series also act (differently) on 
non-series, but not all.

I thought that to qualify to be a series, a type also needed to support 
insertion, not only indexing and content.


cause your above definition qualifies tuple! to be a series, but 
they aren't.

Position: HEAD, TAIL, HEAD?, TAIL?, AT, SKIP, NEXT, BACK

Contents: PICK, POKE (for modifiable series), ordinals (FIRST, ..., 
LAST)

Sequence: PICK and POKE can use numbers (used to implement ordinals)

Persistent position and sequence: All of the above are repeatable, 
with no position change side effect.

LENGTH? too. Note that in R3 an image! counts as a series, though 
it has 2 different position measures.

Ports in R3 (or open/direct or command ports in R2) don't have persistent 
position and sequence, or reference position. This makes them streams, 
not series.

yes basically, they are always at position 0 and looking at the data, 
implies you remove it, so ports are like quantum series  ;-)

The reference position vs inherent position distinction is important. 
That is why the position functions are seeking procedures for ports, 
while they are pure reference-returning functions for series.

It'd still be nice to have an interface that in some ways behaved 
like R2 ports (I think of it as adapter!) that allow the above series 
functions to operate on customised, abstracted series.  Perhaps I'm 
overlooking a simpler way of addressing the problems I have in mind.

I often find myself thinking about, how to categorize REBOL among 
other programming languages. I think, it's how the language wants 
us to make sentences and not just give commands, one on each line. 
In REBOL, we write:

insert back tail series my-value


and it feels natural. In other languages, you would do something 
like:

int sz = length (series);
Series *ptr = &series[sz - 1];
insert (ptr, my_value);

You could maybe do:

insert ( &series[length (series) - 1], my-value);


but it's just far from being as elegant as in REBOL. So most languages 
are like simple commands all the time, one after each other. With 
REBOL we make sentences. It's like a huge step up from many traditional 
programming languages.

Even with a simple example like the one you gave, I might add a couple 
funcs to make the intent even clearer. e.g.

insert-from-tail: func [
    series [series!] 
    value  
    index  [integer!]
] [
    insert skip tail series negate abs index value
]

insert-before-tail: func [
    series [series!] 
    value  
][
    insert-from-tail series value 1
]

insert-before-tail series my-value

Isn't something like this code already built-in in REBOL and I'm 
just missing it:

copy-to: func [series [series!] arg /local cpd d][
    cpd: make type? series 10
    foreach item series [
        either not-equal? arg item [insert tail cpd item][break]
    ]
    cpd
]

no as that would just put the initial values in the series

And in this case, it's easy to try and see if you like it.

length??: func [series [series! none!]] [
    all [series length? series]
]

Everytime I'm writing (slow, but short) code like 

	>> unique append series 'value

or (very fast, but wordy)
 
	>> any [find series value insert series value]
	>> unless find changes 'weight [insert changes 'weight]


I'm wondering whether there's a nicer way to insert a value into 
a series only if it isn't in there yet. Something in the same line 
as ALTER.

This just reeks like the perfect situation for some Guru O'Brian 
or Gabriele D'Enciclopedia to point out that there's already a REBOL 
native which provides exactly that functionality ;-)


On a unrelated side note, I'm wondering if ALTER is missing an /ONLY 
refinement, too:

	>> alter series: [] [1 2] series
	== [1 2]
	>> alter/only series: [] [1 2] series
	** Script Error: alter has no refinement called only
	** Near: alter/only series: [] [1 2] series


Would that be worth adding to R3, I'm thinking about ticketing it 
as a wish in CureCode?

INCLUDE in R3 is not a global word, in the code im currently writing

	>> include package/changes 'weight


reads very nice. Sadly, it's signature wouldn't be compatible with 
EXCLUDE, which only allows series and sets as it's second argument. 

The two refinements /INCLUDE and /EXCLUDE though would make ALTER 
more usefull.

>> spec-of :remove-each
== [

    {Removes values for each block that returns true. Returns remove 
    count. (Modifies)}

    'word [word! block!] "Word or block of words to set each time (local)"
    data [series!] "The series to traverse"
    body [block!] "Block to evaluate (return TRUE to remove)"
]

It will tell you evaluations and series created, which is a bit more 
cross-platform reliable than time. Don't know how fast your CPU is.

forskip+: func [
        "Like FORSKIP, but with local FIRST? and LAST? support."
        [throw catch]

        'word [word!] {Word set to each position in series and changed as 
        a result}
        skip-num [integer!] "Number of values to skip each time"
        body [block!] "Block to evaluate each time"
        /local orig result
    ][

        if not positive? skip-num [throw make error! join [script invalid-arg] 
        skip-num]
        if not any [
            series? get word
            port? get word
        ] [

            throw make error! {forskip/forall expected word argument to refer 
            to a series or port!}
        ]
        orig: get word
        use [first? last?] [
            first?: true
            last?:  false
            body: bind/copy body 'first?

            while [any [not tail? get word (set word orig false)]] [
                if tail? skip get word skip-num [last?: true]
                set/any 'result do body
                set word skip get word skip-num
                first?: false
                get/any 'result
            ]
        ]
    ]

single?: funco [
	"Returns TRUE if the series length is 1."

 series  [series! port! tuple! bitset! struct!] ; map! object! gob! 
 any-word!
][1 = length? :series]

; Note: Type spec same as LENGTH?, which also supports the extra 
types in R3

I have ALTERNATE and MERGE at the lower levels. The first combines 
two series and returns a new series. The second merges one series 
into another, with a /SKIP refinement. I have TO-SPEC-BLOCK , since 
that's such a useful and common need. I avoided using AS- in the 
past, thinking more standard non-copying coercion funcs would make 
them confusing. Those haven't appeared, so I do use AS- sometimes 
now.

I also often use a FILTER function which is just a wrapper around 
remove-each which does a copy on the source input and adds a NOT 
to the comparison block you provide.


this means you keep data instead of removing it and don't break the 
original series.

Actually, it's so easy to add a copy where you need, that i think 
all the  serie's constructing native functions should update the 
input by default, rather than construncting new ones.
Think About UNION, UNIQUE, INTERSECT etc...

They are handy functions but involve too much overhead on big series.

I would love the /INTO refinement be added to all series manipulators.

This way we can make a huge buffer and reuse it all the time.

much less work for the GC and big speed gains on large series.

Yeah, the position being an attribute of the reference is one of 
the core parts of series behavior in REBOL. In contrast, in R3 ports 
the position is an attribute of the port, not the reference to the 
port; same with OPEN/direct ports in R2.

This means that SKIP is a pure function for series, returning a reference 
to a new position - while it is modifying when applied to ports, 
changing the internal position. Same with the other position functions 
like AT, HEAD and TAIL.

I know such things when writing code, but more in a subconcious way. 
Of course different references to the same block having different 
positions doesn't go together with position being an attribute of 
the block (series) itself. That's of course obvious after only a 
little amount of thinking ;-)

A non-parse solution (for the replace problem) based on the existing 
replace mezz:

replace-each: make function! [
	target [series!] "Series that is being modified"
	values [block!] "Block of search/replace strings"
	/local len pos
][
	foreach [search replace] values [
		len: length? search
		while [pos: find target search] [
			target: change/part pos replace len 
		]
	]
]

As long as the first argument is a series ...

Series, or map or object.

its a shame cause looping a large list manually and comparing with 
same?  is VERY slow  :-(

on a 10 million sized series. 


my fastest FIND-SAME loop is 20 time slower than FIND, which would 
be much faster, since all it would have to do is compare a pointer.

here's my fastest one yet... I tried two so far...

	;-----------------
	;-     find-same()
	;-----------------
	find-same: func [
		series [block!]
		item [series! object!]
		/local s i
	][
		i: 1
		repeat s series [
			i: i + 1
			if same? s item [
				break
			]
		]
		unless i > length? series [
			at series i - 1
		]
	]
	

using until was slower probably because REPEAT is faster than UNTIL 
(no exit condition or series index to increase).

How about combining find with a loop
find-same: func [
	series [block!]
	item [series! ]
	/local s 
][
	while [s: find series item] [
		if same? first s item [return  true]
		series: next s
	]
	false
]

You mean using /skip and /compare to sort a series in sets?
s: [1 2 8 a   1 2 6 b   1 2 7 c]    ;; three sets of four items
sort/skip/compare s 4 [3]        ;; sort on the third item
== [1 2 6 b 1 2 7 c 1 2 8 a]

since tuple are not series in R3, then I'd say it's fixed.

they aren't series.

ok, one last one for my rusty rebol

FINDing the index of all occurences of an integer in a series ie:
blk: [ 239 4545 23 655 23 656[
search for 23 and return [3 5]

(not using foreach)

indices?-1: func [
	series [series!]
	value
	/local result
] [
	result: make series 0
	while [series: find series value] [
		append result index? series
		series: next series
	]
	result
]

indices?-2: func [
	series [series!]
	value
	/local result
] [
	result: make series 0
	parse series [

  any [1 1 value series: (append result subtract index? series 1) | 
  skip]
	]
	result
]
>> time-block [indices?-1 blk 23] 0,05
== 0.000006591796875

>> time-block [indices?-2 blk 23] 0,05
== 0.000005645751953125

Hmm, my version is much slower:

indices?-3: func [
        series [block!]
        value  [integer!] 
        /local result
][
        result: make series 0
        parse series [

               any [s: set wanted integer! (if value = wanted [append result index? 
               s])]
        ]
        result
]


>> time-block [indices?-1 blk 23] 0,05
== 0.000004669189453125

>> time-block [indices?-2 blk 23] 0,05
== 0.0000039520263671875

>> time-block [indices?-3 blk 23] 0,05
== 0.0000059661865234375

A bit adapted Lad's version (does not use substraction):

indices?-4: func [
	series [series!]
	value
	/local result
] [
	result: make series 0
	parse series [
		any [series: 1 1 value (append result index? series) | skip]
	]
	result
]


>> time-block [indices?-2 blk 23] 0,05
== 0.000003936767578125

>> time-block [indices?-4 blk 23] 0,05
== 0.000003753662109375

Replacing "1 1 value" by "quote (value)" as variant 5:

indices?-5: func [
	series [series!]
	value
	/local result
] [
	result: make series 0
	parse series [

  any [series: quote (value) (append result index? series) | skip]
	]
	result
]

>> time-block [indices?-1 blk 23] 0,05
== 0.0000047607421875

>> time-block [indices?-2 blk 23] 0,05
== 0.000003936767578125

>> time-block [indices?-3 blk 23] 0,05
== 0.0000059814453125

>> time-block [indices?-4 blk 23] 0,05
== 0.00000372314453125

>> time-block [indices?-5 blk 23] 0,05
== 0.000003692626953125

>> a

== [23 43 55 28 345 99 320 48 22 23 95 884 1000000 999999 999998 
999997 999996 999995 999994 999993 999992 999991 999990 999989 999...
>>

(a is a block with 1, 000,012  integers) 


ind: func[series value x][


	st: now/time/precise
		
	while [x > 0][
		
		result: make series 0
		series: head series

  parse series[any [series: 1 1 value (append result index? series) 
  | skip]] 
		x: x - 1
	]
	et: now/time/precise
	fin: et - st
	print fin
	result 
]

feach: func[series value x][

	result: make series 0 
	st: now/time/precise
	
	while [x > 0][

  foreach[s p v] series [if s = value [append result reduce [s p v]]]
		x: x - 1
	]
	et: now/time/precise
	fin: et - st
	print fin
	result 
]

>> ind a 23 10
0:00:01.249
== [1 10 999990]
>>

>> feach a 23 10
0:00:01.01

== [23 43 55 23 95 884 23 43 55 23 95 884 23 43 55 23 95 884 23 43 
55 23 95 884 23 43 55 23 95 884 23 43 55 23 95 884 23 43 55 23 9...
>>

10 iterations each.. 


foreach is the winner speed wise.. as a bonus, If i use foreach, 
I don't need the index?

in your tests, this causes your loops to slow down a lot:

result: make series 0

you should:

result: make series length? series. 


because when appending, you will be re-allocating the result over 
and over... and the GC is the loop killer in every single big dataset 
tests I've done.  not because its bad, per se, but because you are 
forcing it to operate.

moving the result: make series 0 out of the while loop had a 10 milli 
improvement over 10 iterations.

find-fast: func [
	series
	value
	record-length
	i "iterations"
	/local result st s
][
	
	st: now/precise
	while [i > 0][
		result: make series 0 ;length? series
		s: head series
		until [
			not if s: find  s value [
				either 0 = mod -1 + (index? s) record-length [
					insert tail result copy/part series record-length
					not tail? s: next s
				][
					s: next s
				]
			]
		]
		i: i - 1
	]
	
	print difference now/precise st
	print length? result
	
	result
]

indices?-maxim-1: func [series value /local result][
    result: make block! length? series
    until [
        not if series: find series value [
            append result index? series
            series: next series
        ]
    ]
    result
]

not if series: find/tail series value [
            append result (index? series) - 1
        ]

indices?-3: func [series value /local result][
    result: make block! length? series
    until [
        not all [
            series: find/tail series value
            append result (index? series) - 1
        ]
    ]
    result
]

the kernel of the C extension:

    result = RXI_MAKE_BLOCK(series_n - series_i);
    result_n = 0;
    for (; series_i < series_n; ++series_i) {
        elem_type = RXI_GET_VALUE(series, series_i, &elem);

        if (elem.int64 == value.int64 && elem_type == value_type) {
            result_v.int64 = series_i + 1;

            RXI_SET_VALUE(result, result_n++, result_v, RXT_INTEGER);
        }
    }

btw - will there be any difference between: 

result: make block! length? series
and
result: make series 0


I mean - e.g. not prealocating large enough series (make series 0) 
will be slowed down by GC constantly increasing the size?

there are many important differences:

*make series 0 does not necessarily make a block

*regarding the allocation length - if you can estimate reliably the 
necessary length, then you are better off

what's the quickest way to determine whether a series is past end 
in R2 and then handle it?

As a side note, I noticed, TAIL? and EMPTY? is the same. So you can 
also write the above as:

and series? s empty? s


Is it wise, that TAIL? and EMPTY? does the same? Maybe EMPTY? should 
return false, if the series holds elements, even if you're at the 
tail.

Henrik, can you post some code that will make a series reference 
that is past the end of a series? I'm having trouble making one that 
works consistently, and neither your INDEX? method or EMPTY? will 
work. REBGL keeps adjusting the index returned by INDEX?.

In R3 they stay consistent, and there is an action PAST? to tell 
whether an index is past the end.
>> b: next a: "a"
== ""
>> remove a
== ""
>> index? b
== 2  ; not adjusted
>> past? b
== true


In R2 I don't even see how to write this in mezzanine without temporarily 
appending something on the end of the series (a single element will 
do) and seeing if the index adjusts, then removing what you appended. 
Like this:
>> b: next next next a: "abc"
== ""
>> clear a
== ""
>> also (index? b) < (append b "1" index? b) clear back tail b
== true  ; past end
>> b: ""
== ""
>> also (index? b) < (append b "1" index? b) clear back tail b

== false  ; not past end, but Henrik's method, and EMPTY? will both 
return true.

Here's the source to the backport:
past?: func [
	"Returns TRUE if a series index is past its tail."
	series [series!] ; gob! port!
][
	also (index? :series) < (insert tail :series "1" index? :series)
		clear back tail :series  ; Undo the modification
]

; Note: Native in R3, and non-modifying. No ports because you can't 
undo the

; insert. INDEX? doesn't stay consistent with past-tail references 
in R2.

I'm not sure I would want the series to be modified as the case where 
PAST? is true could be for an indication of the series in a debug 
UI, rather than the actual use of the series. Interesting that you 
had so much trouble producing the problem.

You are also removing an element before the index, but not from the 
beginning. Apparently clearing from somewhere other then the beginning 
of the series breaks the auto-adjusting (which shouldn't be happening 
anyways). PAST? works the same either way.

past?: func [
	"Returns TRUE if a series index is past its tail."
	series [series! gob! port!]
][
	(index? :series) = (index? back :series)
]

; Note: INDEX? doesn't stay consistent with past-tail references 
in R2.

correction:

past?: func [
	"Returns TRUE if a series index is past its tail."
	series [series! gob! port!]
][
	and~ not same? head? :series :series
		(index? :series) = (index? back :series)
]

; Note: INDEX? doesn't stay consistent with past-tail references 
in R2.

Yet another variant:

past?: func [
	"Returns TRUE if a series index is past its tail."
	series [series! gob! port!]
][
	and~ (index? :series) = (length? head :series)
		not same? tail? :series :series
]

; Note: INDEX? doesn't stay consistent with past-tail references 
in R2.

past?: func [
	"Returns TRUE if a series index is past its tail."
	series [series! gob! port!]
][
	and~ (index? :series) = (length? head :series)
		not same? tail :series :series
]

; Note: INDEX? doesn't stay consistent with past-tail references 
in R2.

Simplification:

past?: func [
	"Returns TRUE if a series index is past its tail."
	series [series! gob! port!]
][
	not same? :series skip :series 0
]

; Note: INDEX? doesn't stay consistent with past-tail references 
in R2.

To get a past-tail reference that auto-adjusts, clear from the head 
of the series rather than from further along.

every past-tail series auto-adjusts in ceratain interpreter versions 
(in other no past-tail series auto-adjusts), no exceptions

Henrik does not know any series that does not auto-adjust, it is 
just your error

In 2.7.7, indexes auto-adjust if you clear the series from the beginning, 
but not if you clear from later on. In the same interpreter.

You are getting the same results you got before, it was just your 
error, resulting from the fact, that auto-adjustment adjusts past-tail 
series to tail indices, i.e. to 2 in case Henrik posted

E.g. this modification works regardless of auto-adjustment: (i.e. 
even in R3)

past?: func [
	"Returns TRUE if a series is past its tail."
	series [series! gob! port!]
] [
	and~ greater-or-equal? index? :series index? tail :series
	    not same? :series tail :series
]

BTW, "returns true if a series index is past its tail" mixes different 
datatypes - series index is integer, while tail is a series, so, 
my opinion is, that the shorter wording is less misleading

The not same? :series skip :series 0 line works well enough, afaict. 
Do you have a case for which it doesn't work and the larger test 
is required?

Yes, but skip :series 0 auto-adjusts still, but, as I said, I do 
not mind

(comparing series index - an integer, to series tail - a series)

OK, nevermind, I do not want the text to be longer, a "series is 
past its tail" looks better to me, than any reference to series index

, and it relates to the problem better, since in the case below:

>> b
== ** Script Error: Out of range or past end


what is past tail is just the series not any "series index", which 
should be an integer value, and as such it cannot be "past tail"

I adjusted the http://en.wikibooks.org/wiki/REBOL_Programming/Language_Features/Series
article to mention the past-tail series and to explain differences 
between series with fast indexed access and series with fast insert/remove 
operations.

thing is load isn't the counterpart to mold.  'DO is.   and even 
then, even MOLD/ALL isn't a real serialization because shared objects, 
or series aren't shared anymore.


although they go a long way, and I have been using them extensively 
in many tools, MOLD & MOLD/ALL are just helpers, destined for advanced 
users who have very special requirements.

we need to build a REAL serialization dialect, one which rebuilds 
the data sharing, even in nested and co-occuring references to series 
and objects.

as usually, all datatypes would need to support this extended format, 
but since it would be managed directly within the native datatype 
handling it would be MUCH faster than if we tried to build some dialect 
for it in REBOL.  especially since the native has much more direct 
access to things like same?, and internal states like series index.

foreach allows you to step through a single series one at a time. 
 So, how would you step thru more than one series at a time?

a: [ n elements ]
b: [ n elements ]

foreach [ ofa ofb ] reduce [ a b ] [

] 

?

Is this helpful?

>> trim none

** Script Error: trim expected series argument of type: series port
** Near: trim none

Actually I wonder why find/any shouldn't work on series like these

>> help find
USAGE:

    FIND series value /part range /only /case /any /with wild /skip size 
    /match /tail /last /reverse

Not sure how /match is going to help you here. It only matches the 
start of the series.

>> find/match ["Aa" "Bb" "Cc"] ["Bb"]
== none
>> find/match ["Aa" "Bb" "Cc"] ["Aa" "Bb"]
== ["Cc"]

Here's the R2 version:
extract: func [
	"Extracts a value from a series at regular intervals."
	[catch]
	series [series!]
	width [integer!] "Size of each entry (the skip)"
	/index "Extract from an offset position"
	pos "The position" [number! logic! block!]
	/default "Use a default value instead of none"

 value "The value to use (will be called each time if a function)"

 /into "Insert into a buffer instead (returns position after insert)"
	output [series!] "The buffer series (modified)"
	/local len val
][

 if zero? width [return any [output make series 0]]  ; To avoid an 
 infinite loop
	len: either positive? width [  ; Length to preallocate
		divide length? series width  ; Forward loop, use length
	][

  divide index? series negate width  ; Backward loop, use position
	]
	unless index [pos: 1]
	either block? pos [

  if empty? pos [return any [output make series 0]] ; Shortcut return
		parse pos [some [number! | logic! | set pos skip (

   throw-error 'script 'expect-set reduce [[number! logic!] type? get/any 
   'pos]
		)]]
		unless into [output: make series len * length? pos]
		if all [not default any-string? output] [value: copy ""]
		; R2 PARSE doesn't work well for binary!, so spoof a string!.
		if binary? series [series: as-string series]
		forskip series width [forall pos [
			if none? set/any 'val pick series pos/1 [set/any 'val value]
			output: insert/only output get/any 'val
		]]
	][
		unless into [output: make series len]
		if all [not default any-string? output] [value: copy ""]
		; R2 PARSE doesn't work well for binary!, so spoof a string!.
		if binary? series [series: as-string series]
		forskip series width [
			if none? set/any 'val pick series pos [set/any 'val value]
			output: insert/only output get/any 'val
		]
	]
	either into [output] [head output]
]