Open GoogleCodeExporter opened 9 years ago
Notice that d1...d4 all stack correctly on each other, but if you turn off one
of the
series by setting its data to null (to keep it in the legend), series after
that point
will restack from zero, causing bars to overlap.
Original comment by mathes....@gmail.com
on 28 Oct 2009 at 10:25
I've fixed this locally -- instead of only looking at the previous series to
see where to build from, this
stores an accumulator, which handles empty series.
Here's the updated jquery.flot.stack.js
(function ($) {
var options = {
series: { stack: null } // or number/string
};
function init(plot) {
// will be built up dynamically as a hash from x-value to
var stackBases = {};
function findMatchingSeries(s, allseries) {
var res = null
for (var i = 0; i < allseries.length; ++i) {
if (s == allseries[i])
break;
if (allseries[i].stack == s.stack)
res = allseries[i];
}
return res;
}
function stackData(plot, s, datapoints) {
if (s.stack == null)
return;
var newPoints = [];
for (var i=0; i < datapoints.points.length; i += 3) {
if (!stackBases[datapoints.points[i]]) {
stackBases[datapoints.points[i]] = 0;
}
// note that the values need to be turned into absolute y-values.
// in other words, if you were to stack (x, y1), (x, y2), and (x, y3),
// (each from different series, which is where stackBases comes in),
// you'd want the new points to be (x, y1, 0), (x, y1+y2, y1), (x, y1+y2+y3, y1+y2)
// generally, (x, thisValue + (base up to this point), (base up to this point))
newPoints[i] = datapoints.points[i];
newPoints[i+1] = datapoints.points[i+1] + stackBases[datapoints.points[i]];
newPoints[i+2] = stackBases[datapoints.points[i]];
stackBases[datapoints.points[i]] += datapoints.points[i+1];
}
datapoints.points = newPoints;
}
plot.hooks.processDatapoints.push(stackData);
}
$.plot.plugins.push({
init: init,
options: options,
name: 'stack',
version: '1.0'
});
})(jQuery);
Original comment by mathes....@gmail.com
on 3 Nov 2009 at 7:34
Oops, I left in findMatchingSeries(), even though it's not needed anymore:
(function ($) {
var options = {
series: { stack: null } // or number/string
};
function init(plot) {
// will be built up dynamically as a hash from x-value to
var stackBases = {};
function stackData(plot, s, datapoints) {
if (s.stack == null)
return;
var newPoints = [];
for (var i=0; i < datapoints.points.length; i += 3) {
if (!stackBases[datapoints.points[i]]) {
stackBases[datapoints.points[i]] = 0;
}
// note that the values need to be turned into absolute y-values.
// in other words, if you were to stack (x, y1), (x, y2), and (x,
y3),
// (each from different series, which is where stackBases comes in),
// you'd want the new points to be (x, y1, 0), (x, y1+y2, y1), (x,
y1+y2+y3, y1+y2)
// generally, (x, thisValue + (base up to this point), + (base up to
this point))
newPoints[i] = datapoints.points[i];
newPoints[i+1] = datapoints.points[i+1] +
stackBases[datapoints.points[i]];
newPoints[i+2] = stackBases[datapoints.points[i]];
stackBases[datapoints.points[i]] += datapoints.points[i+1];
}
datapoints.points = newPoints;
}
plot.hooks.processDatapoints.push(stackData);
}
$.plot.plugins.push({
init: init,
options: options,
name: 'stack',
version: '1.0'
});
})(jQuery);
Original comment by mathes....@gmail.com
on 3 Nov 2009 at 7:35
Thank you so much! :-) I had the exact same issue and your fix works nicely.
Original comment by L0n3.R4n...@gmail.com
on 17 Nov 2009 at 2:01
Hello, based on the last version by mathes, here is a version that can now
stack
horizontally.
bars: { horizontal:true, show:true },
The tick algorithm seems to need some help or update.
Best regards,
Alexandre
http://alexandre.alapetite.fr
Original comment by alexandre.alapetite
on 18 Nov 2009 at 6:17
Line 51 should be replace to allow stack not to be defined:
- if (s.stack == null)
+ if (!s.stack)
Demo:
http://alexandre.alapetite.fr/divers/vrac/20091118-flot-stack-horizontal.html
Best regards,
Alexandre
http://alexandre.alapetite.fr
Original comment by alexandre.alapetite
on 20 Nov 2009 at 4:11
Attachments:
There is another(?) bug where leaving out data from a data set for certain data
points will cause stacking to fail. See attached example.
Original comment by schae...@dfn-cert.de
on 26 Mar 2010 at 12:35
Attachments:
[deleted comment]
problem at
http://groups.google.com/group/flot-graphs/browse_thread/thread/5cd1ea353c37cc39
still remains.
line 333 of jquery.flot.js
Message "s[v] is null"
name "TypeError"
This would have the problem solved:
src="http://flot.googlecode.com/svn/trunk/jquery.flot.js"
src="http://flot.googlecode.com/svn/trunk/jquery.flot.stack.js"
However, if use the jquery.flot.stack.js downloaded from this page, and the
jquery.flot.js released at http://code.google.com/p/flot/downloads/list,
problem remains.
Original comment by yuhongch...@gmail.com
on 21 Apr 2010 at 7:41
hello,
I had to find a way to make the stack plugin works with empty series, and
series that don't contain the same
number of datapoints. So I did a few change to the plugin's code (see attached
patchfile - made against the
trunk). Hope it helps :)
Guillaume
--
Keyade.com
Original comment by guillaum...@gmail.com
on 5 May 2010 at 2:25
Attachments:
@guillaume.conte
Thank you. your changes worked! I had some data points in the middle of the
series that suddenly stopped working.
Your fix introduced a new bug, (or revealed) my graph renders the boarder of
the bar on top even if it has size 0.
a work around is setting lineWidth to 0.
Original comment by Buddy...@gmail.com
on 15 Jul 2010 at 4:39
Attachments:
@alexandre.alapetite
I just went through the code to try to fix what was wrong. I tried your fix and
it worked! and it didn't have the error like my previous post.
Original comment by Buddy...@gmail.com
on 4 Aug 2010 at 6:23
Is this still an issue? And was this patch for flot v 0.6? I applied the
patch but I am still getting a small line for series with data = 0. Is anyone
else having this problem?
Original comment by BrianKim...@gmail.com
on 22 Nov 2010 at 1:38
I've just hit exactly the same problems. The 0.7 changelog indicates that empty
data is now accounted for in the stack plugin, but that does not appear to be
the case. Attaching a small file which demonstrates two problems:
- notice the red on the left is overlapping the green, instead of being drawn
above it, because I removed -4, 0 from the young data
- notice a yellow line drawn even though cram data is 0.
The latter issue is not specific to the stack plugin, and happens every time a
bar with a value of 0 is drawn. Would it be possible to skip the drawing in
that case? Setting fillWidth to 0 works around the issue, but makes the graphs
less pretty.
Original comment by sec...@ichi2.net
on 26 Mar 2011 at 1:53
Attachments:
[deleted comment]
Original comment by dnsch...@gmail.com
on 9 May 2012 at 1:53
The problem still exists in the latest flot...
Original comment by Dieter.V...@gmail.com
on 23 Jun 2012 at 8:42
What worked for me was editing the findMatchingSeries Method on line 50+
(original file):
if (allseries[i].stack == s.stack) {
if(!res) res={};
$.extend(true,res,allseries[i]);
}
My series are pretty small in size so I don't know how the performance will be
with big data.
Original comment by Johannes...@gmail.com
on 17 Mar 2014 at 12:42
none of the fixes work for me. I don't know why. I have the latest 0.8.2 flot
charts. Is there any workaround for this to work ?
Original comment by harishba...@gmail.com
on 3 Apr 2014 at 6:43
Back in 2008/9 when I submitted my patch, I don't know if it got
incorporated or not. I haven't used it in a few years, so am unfortunately
unable to confirm for you.
Original comment by mathes....@gmail.com
on 3 Apr 2014 at 7:28
Sorry about that to ask a follow up here as I am totally new to javascript,
just a beginner. I am trying to produce a stock volume profile chart, e.g.
http://www.analystz.hk/trade/stock-volume-profile.php
is the graph in the middle, with date on x-axis, price chart and horizontal
volume bar with 2 different color solely made by the script above? thanks.
Original comment by fishmkt2...@gmail.com
on 30 Jan 2015 at 4:19
Attachments:
Original issue reported on code.google.com by
mathes....@gmail.com
on 28 Oct 2009 at 10:23Attachments: