Import in Openedge - openedge

Question

Good day.
I have a previous question in this link. On the exported csv, I put on the first line the TABLE NAME. I wish to import this CSV to my system.
My current code is this:
DEF VAR ic as INT.
DEF VAR cTable as CHAR.
INPUT FROM VALUE(SESSION:TEMP-DIRECTORY + "temp.csv").
ic = 0.
REPEAT:
ic = ic + 1.
IF ic > 1 THEN DO:
CREATE cTable.
IMPORT DELIMITER "," cTable.
END.
IMPORT cTable.
END.
INPUT CLOSE.
I know that the code is wrong in CREATE part. How do I do this?
Also, when I EXPORT, there is an additional BLANK line after the last record. How do I remove this without opening the CSV file?

Answer 1

Removing the empty line at the end of your file doesn't fix your problem, worse you will not be able to read the last valid CSV line with the IMPORT statement if you do that (it needs an empty line at the end to work properly).
The actual problem is you get an empty row in your table because the IMPORT DELIMITER "," cTable. fails when the REPEAT block reaches the end of the file to leave the loop it raised the ENDKEY condition. But since you call CREATE cTable. before the loop is left you get en empty entry. I hope this explanation helps you understand how the REPEAT loop works, if you don't know this it looks just like an endless loop without any break condition.
Anyway to fix that problem you can either delete the empty row (like you did before), that is perfectly valid or you can omit the NO-UNDO from the temp table definition, because then the REPEAT will UNDO the CREATE by default.
To your other question about the CSV header line, you have to read the line somehow, I don't think there is a statement to just skip it and start reading at the 2. line in the file.
if you need the header names you can simply define a char variable for every column and import it like:
IMPORT DELIMITER "," cColumn1 cColumn2. /* for every column */
or if you just want to read and ignore it you can use
IMPORT UNFORMATTED cTemp.
with a temp variable that reads the whole line.

Answer 2

If you're trying to import an unknown file format, you might try reading in the entire line of data and then pick through the individual fields. As for question 2, it's better to handle the incorrect data programmatically, rather than trying to change the file itself.
This code reads in each line from the file, skipping any that are blank or null. Then it goes through each comma-separated field in the line and displays them.
DEFINE VARIABLE cTable AS CHARACTER NO-UNDO.
DEFINE VARIABLE cField AS CHARACTER NO-UNDO.
DEFINE VARIABLE iLoop AS INTEGER NO-UNDO.
INPUT FROM VALUE(SESSION:TEMP-DIRECTORY + "temp.csv").
REPEAT ON ERROR UNDO, NEXT:
IMPORT UNFORMATTED cTable. /* Read an entire line from the file. */
IF cTable = "" OR cTable = ? THEN NEXT. /* Skip blank lines. */
DO iLoop = 1 TO NUM-ENTRIES(cTable, ","): /* Break up the line by the comma delimiter. */
cField = ENTRY(iLoop, cTable).
MESSAGE "Field " + STRING(iloop) + ": " + cField VIEW-AS ALERT-BOX.
END.
END.
INPUT CLOSE.
Since the file layout is unknown, all of the fields are read as character. You'll need to add some logic to determine if the values are integers, decimals, dates, etc.