Generating rows for missing date

[Coco Qu]

I want to do the interpolation for last copper price. However, original data has missing rows for date. I would want to first fill in rows for missing date with other variables with value 0 and then do the interpolation, but I do not know how to do it.

The date looks like this:
TransactionDate last closingbid closingask
1/1/1904
1/2/1904 6 6.25
1/4/1904 5 7
1/5/1904 7 5.5 6.25
1/6/1904 6 5.75 6.5
1/8/1904 5.75 7
1/9/1904 6 9
1/9/1904 6
1/11/1904 6.75 8.5
1/12/1904 6.25 6 8
1/13/1904 6.25 8
1/14/1904 6.25 8
1/15/1904 6.25 8
1/16/1904 6.25
1/18/1904 6.25 8
1/19/1904 8 6.25 8
1/20/1904 6.25 9
1/21/1904 6 9
1/22/1904 6 9
1/23/1904 8
1/25/1904 6 9
1/26/1904 6 9
1/27/1904 7 10
1/28/1904 7 10
1/29/1904 7 10
1/30/1904 7 10

[Hemanshu Kumar]

You should tsset your data, and then just use the command tsfill. This assumes your TransactionDate variable is already in a numeric form that Stata understands as a date. If not, you will need to create such a variable first. E.g. if it is a string variable, you will need to use the daily() function. Check the Stata help for these, if they are new to you.

[Nick Cox]

Various questions:

1. Have you really got data for 1904? Sometimes there is a translation problem because other software calculates daily dates relative to a zero in 1900.

2. There seems to one day of the week that is never present in the data. Contrary to #1, if the dates are taken literally, they were Sundays.

: absent = mdy(1, (3, 10, 17, 24), 1904)

: absent
            1        2        3        4
    +-------------------------------------+
  1 |  -20452   -20445   -20438   -20431  |
    +-------------------------------------+

: dow(absent)
       1   2   3   4
    +-----------------+
  1 |  0   0   0   0  |
    +-----------------+

:

This situation calls for a business calendar, not the invention of what would be 1/7 of your dataset, with a spuriously inflated sample size.

3. Duplicates can be seen in the data example, which would prohibit the use of tsset until fixed.

[Coco Qu]

Thank you so much!