On 2017-04-20 11:13, Ted Roche wrote:
A parsing routine is usually the fastest. LLFF's (don't forget to use _MLINE) or FileToStr() or Append to Memo and then parse each come out about even in terms of processing time and code, unless the file is really extreme ( 10 Gb, or million byte lines or funny characters...)
Whil ran into a situation where long fields and an impossible number of columns made that technique impractical, and he came up with an innovative solution: import the datea into SQLIte, then use VFP ODBC to manipulate the date from there. He published his whitepaper here: http://hentzenwerke.com/catalog/sqlite2gb.htm
[Disclaimer: I tech-edited this, but make no money from promoting it.]
I do something similar--I use MariaDB/MySQL to import the text file to a table there, having all fields be TEXT fields. Works great. Why reinvent the wheel?