17-Peridot

Solved

Weird reverse

Forum|Forum|11 years ago
August 27, 2014
4 replies
5740 views

Hello folks,

I do still have an unexpected result from the reverse function although it works as expected in some instances; see attached sheet.

When creating an array and reversing the order of the second column then one sees in the graph the reversed function. I.e. last values being depicted first.

However, when I do the same on some measurement data, all I get is the weird looking curve in the left graph.

Does anybody have an explanation? Mike perhaps?

Raiko

1_reverse_Rev2.zip

Statistics_Analysis

Best answer by Werner_E

Mike Armstrong wrote:
Well put Werner, better explanation than I managed

We can see the effect better if we thin out the data to, lets say just 100 data points.

M

MikeArmstrong

1-Visitor

It seems to me that your graphs are plotting from right to left?

When you reverse the data the extreme values of rawdata<1> correspond to smaller values of rawdata<0>. To me the results are what I would expect.

R

Raiko

Author

17-Peridot

Sorry Mike, I mixed up the zip files. Here we go with a new one.

What I don't understand is why the graph of the reversed measurement points is not the mirror image of the raw data as shown in the right hand graph.

Raiko

1_Reverse_2.zip

W

Werner_E

25-Diamond I

Raiko,

I also expected a mirrored graph at first until I realized that this woul only happen if the abscissa values are equally spaced as they are in your test plot to the right.

But they aren't in your measurement data.

Here is a plot with your data thinned out to 100 points only. You can clearly see that we have just about half a dozen higher ordinate values. These are stretched out to the right in the nonreversed plot but crashed together to the left in the reversed one.

To get a mirrored plot you can either manually reverse the limits on the abscissa axis as Mike had done or you create a new vector for the absicissa with 150-data<0> where 150 is the maximum of data<0>.

Embedding pictures does not fork for me at the moment (this happens ever so often here) so I attach them.

1_intriguing-reversal_WE_1-xmcdz.zip

M

MikeArmstrong

1-Visitor

Are you looking for something like this?

2_intriguing-reversa_MA_01-xmcd.zip

1_measurement.xlsx

W

Werner_E

25-Diamond I

Mike,

you don't change anything if you reverse both vectors.

You have changed the limits on the absicissa manually and if you do that in the original plot you get the same efect 😉

M

MikeArmstrong

1-Visitor

Werner Exinger wrote:
Mike,
you don't change anything if you reverse both vectors.
You have changed the limits on the absicissa manually and if you do that in the original plot you get the same efect 😉

Oops yes. The graphs are working correctly IMO.

W

Werner_E

25-Diamond I

I agree with Mike - its all OK.

The puzzling effect occurs because the values on the abscissa are not equally spaced.

If you reverse one of the columns, the higher ordinate values get related to very narrow spaced abscissa values which looks like a spike.