Read 6 bytes floating from a file

[Miguel Pinheiro]

Hello,

I have an old fashion Database with several random access files and I need to import the data to a new Sql Server dababase the data.

There are tables with 6 Bytes floating numbers fields and we have no Number data type with that length in VB.

I tried and tried and... and can't find.

I found this code in the ERP folder, is AlphaBasic. It use an envoirement namend A-shell from Microsabio .

MAP1 OOCUM'4750		! Cabecalhos de Documentos
  MAP2 SIT'4750,B,1	! Ocupado (Sim/Nao)
  MAP2 KEY2'4750		! Chave Secundaria
   MAP3 FCA'4750,S,4  	! Cliente/Fornecedor/Armazem
   MAP3 KEY'4750		! Chave Primaria
     MAP4 SERIE'4750,S,1	! Serie
     MAP4 TIPDOC'4750,S,1	! Tipo
     MAP4 NDOC'4750,S,5		! Numero
  MAP2 ARMZ'4750,S,3	! Armazem
  MAP2 NRINT'4750,B,3	! Numero Interno
  MAP2 SER'ORIG'4750,S,1	! Documento Origem
  MAP2 NUM'ORIG'4750,B,2	! Documento Origem
  MAP2 DATORIG'4750,S,8	! Data Documento Origem	
  MAP2 DATREC'4750,S,8	! Data Computador/Prazo 
  MAP2 DATDOC'4750,S,8	! Data Documento

* * * Fields I Can't convert:

  MAP2 VALDES'4750,F,6	! Valor Desconto
  MAP2 VALPOR'4750,F,6	! Valor Portes
  MAP2 VALSEG'4750,F,6	! Valor Seguro
  MAP2 VALIMP'4750,F,6	! Valor IVA
  MAP2 VALDOC'4750,F,6	! Valor Iliquido

* * * 

  MAP2 TRANSP'4750,S,1	! Codigo Transporte
  MAP2 FLAGEMI'4750,S,1	! Emitido (Sim/Nao)
  MAP2 COD'PAG'4750,B,1	! Codigo Condicoes Pagamento
  MAP2 VENDOR'4750,S,3    	! Vendedor
  MAP2 FNDOC'4750,S,6	! Documento Referencia
  MAP2 DATFACT'4750,S,8	! Data Documento Referencia
  MAP2 NMOV'4750,B,1	! Linhas Documento

* * *

  MAP2 PAGO'4750,F,6	! 

* * *
  MAP2 RESP'4750,B,1	! Linhas Satisfeitas
  MAP2 MOVANUL'4750,B,1	! Linhas Anuladas
  MAP2 ANULADO'4750,S,1	! Anulado (Sim/Nao)
  MAP2 FILLER'4750,S,23

The Fields I Can't convert are Floating with 6 Bytes

Thank you,

Miguel

[R MacDonald]

Hello, Miguel,

I have not done this in VB, but you may be able to use the BitConverter class to accomplish what you need. You will need to know how the 6-byte floating values are structured. This includes which byte is the most significant and which the least, and also which bits within the 48 bit array comprise the mantissa and which bits make up the exponent.

You will need to move the 6 bytes (in the correct order) into a 64 bit Integer and then shift the exponent bits into the proper location for a Double precision floating value. You will need to extend the sign bit of the mantissa into the bits that have been "vacated" by shifting the exponent. You will also need to extend the sign bit of the exponent. Then use the Int64BitsToDoub le method of the BitConverter class to get a double precision floating value that VB can use.

You might find the following reference helpful as you work on this project:

http://www.cse.msu.edu/~cse320/Documents/FloatingPoint.pdf

Cheers,
Randy