A survey on ALLTRIM as a Pain in the neck

[SMARTII]

I'm new as you know, but I have never had any program platform in recent memory pad character elements by default.
I can't even remember when I would want a padded character or numeric value.

So, I'm wondering why the default of most character strings such as fields and str(###) were ever set as padded?

Is it true that the default padding on most every occasion of string values or numerics (left-padded) is a Pain in the neck?
If so, is ALLtrim(), Ltrim() Rtrim() the most used functions in A5?
If so, why does A5 keep this default format?
What am I missing?
Does an application bulging with xxxtrim()'s slow it down?

[Stan Mathews]

Re: A survey on ALLTRIM as a Pain in the neck

why the default of most character strings such as fields and str(###) were ever set as padded?

Has nothing to do with padding per se by Alpha. The table structure used the the Foxpro (2.6 I believe) format. That format stores the data in fixed width fields.

I suppose a complete rewrite of the desktop portion of Alpha to use some other storage basis is possible, but not likely.

[Tom Cone Jr]

Re: A survey on ALLTRIM as a Pain in the neck

The Foxpro format Alpha Five uses is a direct descendant of an earlier format created by Ashton Tate for a product called dBase, one of the first widely available relational database systems. DBF files have been using fixed width fields 25 or 30 years. Fixed width fields mean fixed width records. Understanding the ramifications of this is part of the learning curve with Alpha Five and any other product that uses fixed width records.

[Tim Kiebert]

Re: A survey on ALLTRIM as a Pain in the neck

One can also use the - operator to trim trailing blanks. See this page in the help
I have never had a problem using it. Not sure on the performance issue. (another one of those things I need to test with Ira's utility).
The reason str() is padded on the left is - from the help.

STR() converts a Number to a right-justified character string.

To get a non left padded character from a numeric you can also use Alpha's automagical type conversion as described here.
eg

?str(123)
= "       123"

?""+123
= "123"

?"Qty is " + 123
= "Qty is 123"

I realize this is a v9 forum but here is a quote from the Whats new in v10 about a new xbasic method.

.Data() Method On DBF Tables and Property Variables

When using AlphaDAO, the method used to read data from a resultset is <resultset>.data("columnname")
In order to make it easier to write Xbasic code that will work the same when iterating over a DBF table, or a SQL resultset, or a property variable, the .data() method is now supported for both DBF tables and property variables.
For example:
t = table.open("customer")
?t.lastname
= "Smith "
?t.data("lastname")
= "Smith"

Notice that the <tbl>.data() method trims trailing blanks on character fields.

dim p as p
p.firstname = "Frank"
?p.data("firstname")
= "Frank"

The example using Smith actually has many more trailing blanks but the forum quote feature trimmed them (ironic), but you get the idea.

[Tim Kiebert]

Re: A survey on ALLTRIM as a Pain in the neck

Here are some times courtesy of Ira's Code Utility

ltrim(str(123))   '.000075117371 seconds
""+123            '.000077190179 seconds

alltrim("aaa   ") '.000062034381 seconds
"aaa   " - ""     '.000061815691 seconds

As you can see there is not much in it.

[SMARTII]

Re: A survey on ALLTRIM as a Pain in the neck

OK,
So, instead of:
?rtrim(CUSTOMER->ADDR) + ltrim(str(208))
= "PO BOX 96208"

I just use:
?CUSTOMER->ADDR - "" +208
= "PO BOX 96208"

Trivia: A child of InFormix
What I was used to was:
?CUSTOMER.ADDR | str(208) (no space bet.)
= "PO BOX 96208"

and:
?CUSTOMER.ADDR & str(208) (1 space bet.)
= "PO BOX 96 208"

[SMARTII]

Re: A survey on ALLTRIM as a Pain in the neck

Ok, while you're at it:
Try raw "abc"+str(123)

and raw:
Yourtable->any_char_Fld + str(123)
Even though you get the pads.

Here are some times courtesy of Ira's Code Utility

ltrim(str(123))   '.000075117371 seconds
""+123            '.000077190179 seconds
 
alltrim("aaa   ") '.000062034381 seconds
"aaa   " - ""     '.000061815691 seconds

As you can see there is not much in it.

[Tim Kiebert]

Re: A survey on ALLTRIM as a Pain in the neck

raw "abc"+str(123) was .000076824565

I could not run Yourtable->any_char_Fld + str(123). For some reason the utility did not like accessing a table that way.

But

t=table.open("customer")
t.persist()

?t.LASTNAME + str(123)   [B][COLOR=Blue] 'and only timing this line was .00018009446[/COLOR][/B]
= "Graham                     123"

[SMARTII]

Re: A survey on ALLTRIM as a Pain in the neck

RE:
"..from the What's new in v10 about a new xbasic method."...
Tim,
Thanks!!!!
Because, I tried it in V9:
Ex:
FUNCTION notpadded AS C ()
t = table.open("CUSTOMER")
notpadded = t.data("ADDR")+125

END FUNCTION
---------------
?notpadded()
= "PO BOX 96125"

One More:
FUNCTION notpadded AS C ()
t = table.open("CUSTOMER")
notpadded = t.data("Name") + ", " + t.data("city") + ", " + t.data("zip")
END FUNCTION

?notpadded()
= "1ST AMERICAN BANK TEXAS, AZLE, 76098"

WELCOME TO V9.10

One can also use the - operator to trim trailing blanks. See this page in the help
I have never had a problem using it. Not sure on the performance issue. (another one of those things I need to test with Ira's utility).
The reason str() is padded on the left is - from the help.To get a non left padded character from a numeric you can also use Alpha's automagical type conversion as described here.
eg

?str(123)
= "       123"
 
?""+123
= "123"
 
?"Qty is " + 123
= "Qty is 123"

I realize this is a v9 forum but here is a quote from the Whats new in v10 about a new xbasic method.The example using Smith actually has many more trailing blanks but the forum quote feature trimmed them (ironic), but you get the idea.

[Tim Kiebert]

Re: A survey on ALLTRIM as a Pain in the neck

Well what do you know. :)

I found this in the patch realease notes for v9

<Tbl>.data() Method. Table objects now have a new method to read data from the table. <tbl>.data(fieldname) gets the value of the specified field from the current record.
t = table.open("customer")
ui_msg_box("",t.data("firstname"))

'The above code is a shortcut for this:
t = table.open("customer")
f = t.field_get("firstname")
ui_msg_box("",f.value_get())

And on finding this my ailing (failing) memory thinks it may have seen it there first. :o
I guess may be the new bit for v10 is that it is also supported for property variables. But may be you can test that as well and prove me wrong . And then you will have a v9.10. :)

[SMARTII]

Re: A survey on ALLTRIM as a Pain in the neck

Just the opposite of what you might expect??
No formatting = Longer?
Or did you try formatting from a table field?

raw "abc"+str(123) was .000076824565

I could not run Yourtable->any_char_Fld + str(123). For some reason the utility did not like accessing a table that way.

But

t=table.open("customer")
t.persist()
 
?t.LASTNAME + str(123)   [B][COLOR=blue]'and only timing this line was .00018009446[/COLOR][/B]
= "Graham                     123"

[Tim Kiebert]

Re: A survey on ALLTRIM as a Pain in the neck

This is reading from a table which requires a disk read ( or maybe at least a cache read)
t.LASTNAME + str(123) was .00018009446
which should be pretty much the same as - Yourtable->any_char_Fld + str(123)

And this is not
"abc"+str(123) was .000076824565

[SMARTII]

Re: A survey on ALLTRIM as a Pain in the neck

I have nothing in my memory..failing or otherwise..
So,
Give me an example that would qualify.
ps.
Just so you know:
data() is nowhere to be found in the V9 HELP (Search or Index).
Of course same for my V10 Help..nowhere..

Well what do you know. :)

I found this in the patch realease notes for v9
And on finding this my ailing (failing) memory thinks it may have seen it there first. :o
I guess may be the new bit for v10 is that it is also supported for property variables. But may be you can test that as well and prove me wrong . And then you will have a v9.10. :)

[MikeC]

Re: A survey on ALLTRIM as a Pain in the neck

data() is nowhere to be found in the V9 HELP

Not in the help files (reason why Tim looked in the release notes and What's New areas).... but does come up in the code/interactive editors.

Help for parts of v8, v9, v10 are now in those 3 places....4 counting this messageboard!

[FrankR]

Re: A survey on ALLTRIM as a Pain in the neck

Give Alpha + SQL Server Express a try.