Reading plain text IN dos ? Funny input!

[Robert E. Carneal]

I am attempting to read a text file using PowerBasic DOS. I have checked the PowerBasic manuals, searched online, and I am positive the answer is staring at me in the face, but I am completely missing it!! :SHOCKED: Here are three lines from the data file for example:
"358293","Pearl and Hermes Atoll","Island","HI","15","Honolulu","003","275000N","1755000W","27.8333333","-175.8333333","","","","","0","Unknown","09/30/2003",
"358294","Laysan Island","Island","HI","15","Honolulu","003","254615N","1714415W","25.7708333","-171.7375","","","","","3","Unknown","09/30/2003",""
"358295","Barking Sands","Beach","HI","15","Kauai","007","220418N","1594652W","22.0716667","-159.7811111","","","","","0","Kekaha","02/06/1981",""

(yes, each line is on a single line, but they are three separate lines. I separated the lines using colors so you can tell them apart.)
Ok, they all read with spaces interspersed with spaces. When my program reads this file for example, "Pearl and Hermes Atoll" becomes " P e a r l a n d H e r m e s A t o l l " including the double quotes! Would this be an example of a binary file or anything? This data file contains some characters that have diatribe marks on them. That's ok, right? (I hope so!)

Here is my line that I am reading the above with:
Input #1,FeatId$,FeatNm$,FeatCl$,StatAl$,StatNu$,CntyNa$,CntyNu$,PLaDMS$,PLoDMS$,PLaDEC$,PLoDEC$,SLaDMS$,SLoDMS$,SLaDEC$,SLoDEC$,Elevat$,MapNam$,DateCr$,DateEd$

Only thing I know to do at this point it remove the spaces from each and every variable one at a time, and rerun my program at that point. That would take a couple days- the data file is that large. Does anyone see something stupid I may have done?

Please excuse my use of colors in this post. I was trying to be clear and make sure it was understandable.

Thank you for your understanding, everyone.

Robert

[Mel Bishop]

Looks to me like you are reading a unicode file.

A possible fix:

temp$
for x = 1 to len(rString$) step 2
temp$ = temp$ + mid$(rString$,x,1)
next x

[Robert E. Carneal]

Unicode. <cuss> (IF I say what I am thinking, someone would probably kick me off the site.) First time I ever ran across one. Now that I know, the solution is deceivingly simple. I was sure this was laughing at my face. (But, hey, I never saw a Unicode file before, so can I be forgiven?) Thanks!

Robert

[Mel Bishop]

Unicode...can I be forgiven?)

Sure. Why not? Besides, do you think you are the only one?

[Robert E. Carneal]

temp$
for x = 1 to len(rString$) step 2
temp$ = temp$ + mid$(rString$,x,1)
next x

It works. Since I did that- I found a faster way, but non-Powerbasic. All I had to do was do this:

C:\> type Unicode.txt > file.txt

This removes the Unicode of this. You need to decide if that is what you really want to do first before doing it. I don't know how to reverse this is you do it by accident.

Maybe that will help someone else? I did not know you could do that. I was impressed with the speed of that- it was really quick.

Thanks,

Robert

[Scott Slater]

You could do it like this

Function FixTextLine(TextLine As String) As String
 
   Dim OutString   As String
   Dim i           As Word
   Dim o           As Word
   Dim pIn         As Byte Ptr
   Dim pOut        As Byte Ptr
 
   pIn = StrPtr(TextLine)
   OutString = Space$(Len(TextLine))
   pOut = StrPtr(OutString)
   o = 0
 
   For i = 0 To Len(TextLine)-1
      If @pIn[i] < 127 And @pIn[i] > 0 Then  ' Modify this range as needed or add ELSEIF sections to expand
         @pOut[o] = @pIn[i]
         Incr o
      End If
   Next i
 
   FixTextLine = Left$(OutString,o)
 
End Function
 
Dim LineData As String
 
Open "MYFILE.TXT" For Input As 1
Line Input #1, LineData
Close 1
 
Print "Data as read from the file:"
Print LineData
Print
Print "Fixed Data:"
Print FixTextLine(LineData)
End

Not tested, but it should work.

[Scott Slater]

If they are spaces Chr$(32) then use the following instead.

Function FixUnicodeTextLine(TextLine As String) As String
 
   Dim OutString   As String
   Dim i           As Word
   Dim o           As Word
   Dim pIn         As Byte Ptr
   Dim pOut        As Byte Ptr
 
   pIn = StrPtr(TextLine)
   OutString = Space$(Len(TextLine))
   pOut = StrPtr(OutString)
   o = 0
 
   For i = 0 To Len(TextLine)-1 Step 2
      @pOut[o] = @pIn[i]
      Incr o
   Next i
 
   FixTextLine = Left$(OutString,o)
 
End Function

Pointers should go lots faster than string concatenation and MID$

[Michael Mattias]

C:\> type Unicode.txt > file.txt

Above the Hard/Long way? Or at least "out of your program's control?"

OPEN MyFileFor BINARY AS #6529
GET   #6529, LOF(6529), UnicodeText$ 
RegularText$ = ACODE$(UnicodeText$) 
SEEK 6529, FILEATTR(6529,-2&)
PUT$ 6529, RegularText$
SETEOF #6529
CLOSE  #6529

Not sure it's UNICODE?

Let's add one more step

OPEN MyFileFor BINARY AS #6529
GET   #6529, LOF(6529), UnicodeText$ 
IF ISTRUE (IsTextUnicode (BYVAL STRPTR(UnicodeText$, %NULL))  THEN
     RegularText$ = ACODE$(UnicodeText$)
     SEEK 6529, FILEATTR(6529,-2&)
     PUT$ 6529, RegularText$
     SETEOF #6529
END IF 
CLOSE  #6529

Leaves non-Unicode text alone, converts Unicode to ANSI

MCM

[Michael Mattias]

This is PowerBASIC for MS-DOS isn't it?

Never mind.

MCM

[Scott Slater]

This is PowerBASIC for MS-DOS isn't it?

Never mind.

MCM

Hehe, It's hard to remember what is supported in PB-DOS sometimes after using all of the new stuff in the Windows compilers for so long.

[Michael Mattias]

Part of me is saying "take a few days and work with your MS-DOS compiler. You haven't touched it for five years and you are getting really rusty."

The other part is saying "Don't do that, it would be a waste of time because MS-DOS is dead."

[Manuel Valdes]

I think it's not neccessary to reverse, since you now have two files, the original one and a modified one. Anyway if you need to reverse the modified may be you can open it with the Notepad and use the option 'save as', selecting the unicode format.

[Cliff Nichols]

When dealing with TEXT and not knowing if UniCode or not....I TOTALLY aggree with you

Unicode. <cuss> (IF I say what I am thinking, someone would probably kick me off the site.)

I have wasted DAYYYYYS before (less often lately) before all the sudden it hits me....."WTF??? could there be unprintable characters in this text???"

More often than not, thats what it turns out to be, and just never occurred to me to even check (especially when none of the printable text has any extended characters)

a quick idea to track down would be to search each character and verify it at least is between 0 and 127 and then go from there for characters you do not expect.

[Joseph Cote]

a quick idea to track down would be to search each character and verify it at least is between 0 and 127 and then go from there for characters you do not expect.

Be careful with zero. If it is the default ansi to unicode conversion then the high byte for each character will be zero. Think of it as 8-bit chars converted to 16-bit with leading zero fill.

Also.... From the original post, it looks like some of the island names are using the higher unicode characters. What they will look like when stripped of the high byte may not look at all correct. How to fix that I can't say, just mentioning it as a possible problem.