Re: Sorting paired columns of dates and values
- To: mathgroup at smc.vnet.net
- Subject: [mg106699] Re: Sorting paired columns of dates and values
- From: Garapata <warsaw95826 at mypacks.net>
- Date: Thu, 21 Jan 2010 04:54:29 -0500 (EST)
- References: <hj6qj5$8vc$1@smc.vnet.net>
I've made a bit of progress on this with the following functions:
uniqueDates[data_, year_] := Module[{step1, step2, step3},=E2=80=A8 step1 =
Flatten[Rest[data][[All, Range[1, Dimensions[data][[2]], 2]]]];
step2 = Delete[step1, Position[step1, ""]];=E2=80=A8 step3 = Union[step2];
DeleteCases[step3, _?(ToExpression[StringTake[#, -4]] < year &)]];
This takes my original data as an input and returns a sorted list of
unique dates. I need this (or some functional equivalent) because I
discovered that each of my paired data and value columns may not have
the same set of dates between any two given dates. Some may not have
values for holidays or weekends others may. Also, as I described in
my first post, some of the paired columns start with earlier dates
than others.
I also discovered in the data set that I have "" (empty strings) in
some positions for both dates and data values.
Next I wrote the following:
cleanData[dataIn_, index_] := Module[{temp1},
temp1 = Rest[dataIn[[All, index ;; index + 1]]];
Delete[temp1, Position[temp1[[All, 1]], ""]]
];
The above, takes my original data and an index corresponding to a date
column of one of the paired columns of date and values from the
original data as input. It returns a matrix with just 2 columns
corresponding to dates and values having delated rows without data.
Note: the Rest[] just takes out a header row.
Next:
deleteDatesBeforeYr[data_, year_] :=
DeleteCases[
data, _?(ToExpression[StringTake[#[[1]], -4]] < year &)];
The above, takes input from the output of cleanData[] and deletes rows
with dates prior to a specified year.
These last two functions only work on one date and value paired column
from the original data set at a time, so they could stand some
improvement to work all at once.
Still, at this point I can produce:
1. A list of sorted unique "master" dates and
2. Several matrices each with a date and data column.
Now I need to combine them into a single matrix with the unique
"master" dates in the first column and the data from each of the
several 2 column matrices aligned in the correct row of the date
column.
(I know this will leave me with some blank entries at some positions
in the data columns of the matrix. I need to give some thought to
what to do with these later.)
After these 3 posts and longwinded explanations, if anyone has stayed
with this and can suggest simpler ways to do the above and a way to do
the next step to get me what I need, I will very much appreciate the
guidance.
Thx
G
(Wouldn't it make things a bit easier if we could edit our posts?
Just wondering)