DOCUMENTATION.md

CSVProcessor Documentation

Overview

With CSVProcessor, a CSV file can be imported in a similar way that the IMPORT command imports other file types (XLS, WRK, ...), with some added features.

• Properties and Methods
• Examples

Quick guide

To enable the class

DO LOCFILE("csv.prg")

To instantiate an object

LOCAL CSV AS CSVProcessor

m.CSV = CREATEOBJECT("CSVProcessor")

The properties that control the CSV parsing must match the formatting of the source file. For instance, to import data from a CSV file that is delimited with tabs, the decimal point is represented by a comma, the dates are inserted according to the basic format of ISO 8601 (YYYYMMDD), and empty values represent NULL:

m.CSV.ValueSeparator = CHR(9)
m.CSV.DecimalPoint = ","
m.CSV.DatePattern = "%4Y%2M%2D"
m.CSV.NullValue = ""

To import the CSV data, and store the result in a specific new cursor

IF m.CSV.Import(GETFILE("csv"), "tempCSV") = 0
  SELECT tempCSV
  BROWSE
ELSE
  * an error occurred
ENDIF

If a date column in the CSV file was not properly imported as a date field, it's easy to check for the values that could not be parsed as a date. For instance, assuming that the data was imported into the cursor tempCSV, and the field name is dateCreation , looking for mismatched column values can be done this way:

SELECT dateCreation FROM tempCSV WHERE ISNULL(m.CSV.ScanDate(dateCreation, .F.))

Hopefully, this will help on setting an adequate date pattern. The same type of verification can be done with Logical and Numeric columns.