Import File Size

[David Thompson]

>The only thing I see is tons of empty cells with quote wrappers, and commas
>to delimit fields. Scanning through the BBEdit doc, this visually appears to
>be 20%.

If every cell has a quote wrapper, that would probably account for 
it. The comma separators would just about be a wash. Prior to each 
cell in a Panorama database, there will be one or more bytes to 
indicate the data size for that cell. For example, if there are 20 
characters in the cell, they will be preceded by chr(21). The size 
indicator includes itself in the count. An empty cell would just have 
the chr(1) size indicator and nothing else. I'm not sure exactly 
where it becomes necessary to use more than one byte to indicate the 
size, but I think it's somewhere around 120. Most ordinary fields 
containing names, addresses etc. are small enough that only one byte 
will be needed for that, so that should just about cancel the one 
byte used by the comma delimiter. Each record has a header, and I 
think also a trailer to indicate where the adjacent records are, 
whether the record is selected, its summary level and so on. The 
memory needed for that would probably be equal to that used by the 
quotes in the first 4 to 6 fields in the text file, and the carriage 
return separators. If the text file uses crlf as the line ending, 
that would be one more character per line in the text file. A brand 
new Panorama database, with no procedures, forms, crosstabs, 
permanent variables, or images in the flash art scrapbook would use 
very little memory for those modules. The memory used to hold the 
design sheet data would be related to the number of fields, and not 
to the number of records. A brand new database named "test" with one 
empty record of ten fields is 2054 bytes. With more fields, maybe its 
3000, but that's a tiny percentage of a 100 MB database.

Dave