[R-SIG-Finance] getSymbols.yahoo 'adjusting' to NA

[Joshua Ulrich]

Garrett,

My attempt to patch this a couple weeks ago uncovered some other
issues (via R CMD check) that I wanted to resolve before responding.
As of revision 111 on R-forge, adjRatios() ignores split/dividend data
that do not have corresponding close prices.

Sorry for the delay.

Best,
--
Joshua Ulrich  |  FOSS Trading: www.fosstrading.com



On Mon, May 16, 2011 at 10:43 AM, Joshua Ulrich <josh.m.ulrich at gmail.com> wrote:
> Hi Garrett,
>
> On Sun, May 15, 2011 at 4:33 PM, G See <gsee000 at gmail.com> wrote:
>> Sorry. I see what you mean. adjustOHLC calls getDividends() without 'to'
>> # Try to do, "by hand," the same thing
>> # that getSymbols.yahoo does
>> getSymbols('SPY', to='2010-01-01')
>> div <- getDividends('SPY')
>> spl <- getSplits('SPY')
>> adj <- na.omit(adjRatios(spl, div, Cl(SPY)))
>> #head(adj); tail(adj)
>> fr <- SPY
>> fr[,1] <- fr[,1] * adj[, "Split"] * adj[, "Div"]
>> ###End code
>> Now it hangs here
>> Thanks for looking into this.
>>
> I consider this a bug in TTR::adjRatios.  It should discard any
> split/dividend data that lie outside the range of the close data.  I
> will patch tonight.
>
> This issue manifests in adjustOHLC because all historical
> split/dividend data for "x" is pulled, not just historical data up to
> the last date in "x".  Perhaps adjustOHLC should use a "to=" argument
> for getDividends and getSymbols as well...
>
> Thanks for the report!
>
> Best,
> --
> Joshua Ulrich  |  FOSS Trading: www.fosstrading.com
>
>
>
>> On Sun, May 15, 2011 at 4:13 PM, G See <gsee000 at gmail.com> wrote:
>>>
>>>> Because you included the 'to' argument in your call to getDividends
>>>> and getSplits; and it's not included in those calls within
>>>> getSymbols.yahoo.  I'll take a closer look at this...
>>>>
>>>
>>> I think it is worth mentioning that getSymbols.yahoo does, in fact, use
>>> the 'to' argument to call getDividends and getSplits and is at least one
>>> reason you get big differences in yahoo's Adjusted price and the Close price
>>> after you adjust using either adjustOHLC(), or getSymbols() with
>>> adjust=TRUE.
>>>
>>>>
>>>> Best,
>>>> --
>>>> Joshua Ulrich  |  FOSS Trading: www.fosstrading.com
>>>>
>>>
>>> Garrett
>>
>