Bookmark and Share

Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: st: Interpolating missing prices

From   Nick Cox <>
To   "" <>
Subject   Re: st: Interpolating missing prices
Date   Sun, 24 Jul 2011 12:26:11 +0100

-ipolate- would be natural if you wanted to assume linear change of price with time between non-missing values. Alternatively, there is an FAQ on replacing missings with previous non-missings which includes a reversing time trick for getting following non-missings.

. search missing, FAQ

should you point you to it.


On 24 Jul 2011, at 07:52, Beatrice Crozza <> wrote:

For what I have understood, ipolate interpolates the missing values of
a variable with the values of another one.

In my case, prices are not a function on another variable., so I want
to interpolate only prices.


2011/7/24, Nick Cox <>:
Are you asking for -ipolate-?


On 23 Jul 2011, at 22:45, Beatrice Crozza <>

Dear All,

I have a high-frequency price series with many missing values:

date                     time      price
26jan2008    83000
26jan2008    83500
26jan2008    84000
26jan2008    84500    100.1
26jan2008    85000
26jan2008    85500
26jan2008    90000
26jan2008    90500
26jan2008    91000    100.14
26jan2008    91500
26jan2008    92000    100.12
26jan2008    92500
26jan2008    93000    100.15

I would like to interpolate missing prices with the aritmetic mean of
the previous and the next prices to the missing one.

I typed:
by date: gen price2=(price[_n-1]+price[_n+1])/2 if price[_n]==.

however, I receive many missing values as result.

How can I ask Stata to use for the interpolation only the available
prices, avoiding the missing values?

*   For searches and help try:

© Copyright 1996–2018 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   Site index