I also had this problem. But my advisor told me that it's impossible to handle the holiday problem by Eviews. Usually, there are two ways to solve the holiday problem:

First, you could use linear method to add the holiday missing data. For example, you can make the average by the day before and after. It is reasonable, but it may be bias.

Second, ignoring the holiday data for your dataset, which means deleting all holiday data for other variables. It is resonable for doing econometric analysis, but you need to pay more attention for the date of data.

谢谢！

请问第二种方法如何操作呢？

Use the Matrix in Eview. For example: the sample period is 01/01/2009 to 01/31/2009, which means your should have 22 days'data.

Let's assume you have the daily Stock data and the daily Highest Temperature data.

The Stock data at least miss the new year' data (01/01/2009), so you will have 21 obs for Stock data and 22 obs for daily Highest Temperature data.

Now we have a matrix (22 rows and 5 columns): first column is for the date of Stock data, second column is for the value of Stock data, third column is for the date of Temperature and the forth column is for the value of Temperature.

' assign the group data into a matrix

matrix(22,5) xdata
for !x =1 to 22
xdata(!x,1)=date1(!x)
xdata(!x,2)=value(!x)
xdata(!x,3)=date2(!x)
xdata(!x,4)=temp(!x)
next

' the data in matrix column 1 & 5 is the data you need.
for !i = 1 to @rows(xdata)-1
  if xdata(!i,1) = xdata(!i,3) then
  xdata(!i,5) = xdata(!i,2)
  else  for !j = !i+1 to @rows(xdata)
            if xdata(!i,1) = xdata(!j,3) then
            xdata(!i,5) = xdata(!j,4)
            exitloop
            endif
        next
   endif
next