<html>
<head>
<meta http-equiv="content-type" content="text/html; charset=UTF-8">
</head>
<body text="#000000" bgcolor="#FFFFFF">
Nyall wrote:<br>
<blockquote type="cite"><span>I suspect the lines aren't **exactly**
coincident. I'd try:
</span><br>
<br>
<span>- multiparts to single parts
</span><br>
<span>- snap geometries to layer, to snap the end points exactly
to each other
</span><br>
<span>- retry <span class="bold highlight search-highlight">merge</span>
</span></blockquote>
<br>
I tried this, no joy. still getting multi-linestring with 2055
segments.<br>
So here's what I tried next - <br>
multiparts to singlepart<br>
export to csv, saving geometry as wkt.<br>
verify that endpoints match <i>exactly</i> to the precision written
to the file<br>
Import csv as layer<br>
merge<br>
<br>
again, no luck<br>
<br>
I looked at the attribute table and the lines were in some random
order- adjacent segments were not consecutive rows in the attribute
table. Then I noticed that every row had the same FID. <br>
I closed the layer, opened the .csv in excel, verified that the rows
represented consecutive points, then assigned consecutive FIDs. <br>
imported csv as layer<br>
merge<br>
<br>
again, no luck. Still have 2055 rows in the attribute table of the
merged layer, in that same seemingly random order as before. <br>
<br>
A purely brute force solution was to export as csv, as above, then
edit the file by removing the endpoint of each segment, and
concatenate them into a single linestring with 2055 points. <br>
</body>
</html>