Add -Fe for step-up sampling in sample1d

[PaulWessel]

See the forum post for details. This interpolator keeps the previous value until we pass a new input node and then we step up to that value.

@Esteban82, I think this reproduces your example but I do not know if it is generally correct since I only tested it on yours.

[joa-quim]

Don't replace those spaces by tabs. Using the spaces like that is the only way of keep the alignment with the first character after the parentesis.

[PaulWessel]

Sorry, it was an auto think. I will fix it.

[joa-quim]

Likewise here and below.

[PaulWessel]

I did a few more but dont want to go over all of this manually. As you know I am not in favour of 80character screens so feel these break-the-line adds longer code and more opportunities for tab/space shits.

[joa-quim]

It's has nothing to do about the the line width but only on a good alignment with above line. Tee price of indenting with tabs is that we must do a mix of tabs and spaces to keep that alignment.

[PaulWessel]

If you want to revert some of these then no problem, go right ahead.

[Esteban82]

Great Paul!! I test it. It works. But I found a bug when the input has two records

cat << EOF | gmt sample1d xy.txt -I1 -Fe
1	1
5   4
EOF
Output:
1       1
2       1
3       1
4       1
5       1

[PaulWessel]

OK, I was checking for the first point first but need to do the end first, then first, then the rest. Now works, I think?

[Esteban82]

I found a new issue.

cat << EOF | gmt sample1d -I2 -Fe -Am
1	1
3   3
7   4
EOF
output:
2       1
4       3
6       4   # here I would expect 3.

[Esteban82]

With the GMT_Splines.sh also found some issues. Maybe the sames (see #7490).
Here is the figure.

gmt sample1d t.txt -Fe -T0.5
0       0
0.5     0
1       1
1.5     1
2       1.25
2.5     1.25
3       1.25
3.5     1.25
4       1.5
4.5     2
5       2
5.5     2
6       2.5

[PaulWessel]

-Fe requires monotonously increasing data but this data set is not that, so not an appropriate test

[Esteban82]

Ok, I see. So, would I need to pre process my data with another tool?

[PaulWessel]

Well, it only make sense for cumulative data since we step up

[Esteban82]

In the case of the forum, yes, it is cumulative. But it might not be always the case specially with dates.

[PaulWessel]

Perhaps you can give me some examples of how -Fe would work on regular data such as in GMT_splines.sh. I approached the problem as specific to cumulative data, or at least a monotonically increasing data set, but presumably we could do something similar for monotonically decreasing data. Not sure what expectation is for other data sets though.

[Esteban82]

This is a simple animation that I want to make. Look at the labels.

Messi.mp4

When I interpolate the data (with sample1d -Fe) I got this table.
There are some issues:

• Why the list doesn't start at 19-01 as in the dataset? It does if the first data is 2004-01-18.
• In the last record says 5 ahead of time (14-Feb) when in the data table it says 16-Feb.

times.txt:

2004-01-21T00:00:00	1
2004-01-24T00:00:00	1
2004-01-27T00:00:00	1
2004-01-30T00:00:00	2
2004-02-02T00:00:00	2
2004-02-05T00:00:00	2
2004-02-08T00:00:00	2
2004-02-11T00:00:00	2
2004-02-14T00:00:00	5

Full script

#!/usr/bin/env bash
# Data
cat << EOF > Accumulated_Goals.txt
2004-01-19 1
2004-01-28 2
2004-02-16 5
EOF

gmt sample1d Accumulated_Goals.txt -fT --TIME_UNIT=d -I3 -Fe > times.txt

# 	2. Set up main script
cat << EOF > main.sh
gmt begin
	gmt basemap -Rd -JW24c -Bf -Y0 -X0
gmt end
EOF

gmt movie main.sh -Cfhd -Ttimes.txt -NMessi -Ml,png -Vi -Zs -Fmp4 -Lc0+jTR+o0.3/0.3+gwhite+h+r \
	-Lc1+jTL+o0.3/0.3+gwhite+h+r

[PaulWessel]

I guess 19 is not a multiple of 3, but 21 is. Remember, you are stepping every 3 and nodes are always a multiple of the increment.

[Esteban82]

I guess 19 is not a multiple of 3, but 21 is. Remember, you are stepping every 3 and nodes are always a multiple of the increment.

Ok, thanks, I didn't know that. Altough with dates it doesn't make much sense to have steps that are multiple of the increment.