Change in Pageload characteristics

[econmiller]

I just updated to build Nov 11 2014 on Eviews 8 64bit, and suddenly my programs stopped working. I didn't need to use arguments in the past to read a CSV file with variable names on line 1, but now it automatically skips the header row and gives me series01....series100 instead of using the column names. I tried using wfopen and pageload with and without a combination of arguments including "colhead=1" and "firstobs=0" to no avail. They seem to not change the way the CSV file is loaded.

Steve

[EViews Gareth]

Could you provide one such csv file?

[econmiller]

Here's the statement that used to work

pageload "C:\Users\mill1707\Documents\MI Forecast\Data\bls\emp.csv"

[EViews Gareth]

I went back a number of patches, and back to EViews 7, and the behaviour was exactly the same as the current version.

To get the names in you need:

Code: Select all

pageload emp.csv skip=0

[econmiller]

Really? Either way that worked. Thanks,

[econmiller]

Gareth,

I've attached another file that for all practical purposes looks just like the prior CSV file and it reads it without the skip=0 statement. That is this statement reads it correctly:

pageload "C:\Users\mill1707\Documents\MI Forecast\Data\bls\la.csv"

Can you identify why this works without the option while the emp.csv file does not work?

Also I tried the following statement just for kicks before contacting you using the skip argument. But it did not work. Is there a rule to determine when the arguments should follow the command in parentheses versus at the end of the statement?

pageload(skip=0) "C:\Users\mill1707\Documents\MI Forecast\Data\bls\emp.csv"

[EViews Gareth]

Arguments always follow the command. Options are always in parens. Pageload/Wfopen/Import have few options and many arguments.

EViews always looks at a file and tries to figure out which parts contain things that look like data, which parts contains things that look like row labels (dates/IDs), which parts contains things that look like column labels (series names), and which parts contain other stuff we're not interested in.

In general it does a pretty good job of figuring it out by itself. Sometimes you need to nudge it though. In your first file it looked like the first row was extra crap (probably because only two columns had data immediately following that first row - indicating that the first row was probably some sort of header/title information). In the second file more columns had data immediately after that first row.

Obviously if you always want to be dead certain you know what EViews will do, you can use the full set of arguments to exactly describe the way the file is laid out. The vast majority of the time, though, it is easier to let EViews figure it out and cross your fingers.