+2Hello FME community,
I am quite a beginner with FME and I have point shapefile includes country names. Each country name divided into 2 parts and the 1st part ended with (-).
How can I combine a country name and remove (-) in the following list?
To this:
For each name sometimes 1st part sorted at end, for example Camb-
Does anyone have any ideas or potential solutions?
@takashi @david_r @jlutherthomas @egomm @mark2atsafe
+51Hello @a_afiaax
I simulated here your case with the XLS Reader.
First: filter by attribute Name with symbol " - " - transformer StringSearcher.
Step 2: Create a attribute Order for each results from Output port Matched ( transformer StringSearcher ).
Step 3: Sroter by attribute Order and after Aggregator by attribute Point_ID.
Attached the Workspace Template - Workspace_String.fmwt
Thanks,
Danilo
+60So the general approach would be PointOnPointOverlayer to build the list, ListConcatenator to join the name values and then a StringReplacer to get the - out. But you need to make sure the list is sorted correctly, either by making sure the points enter the PointOnPointOverlayer in the correct order or by sorting it before the ListConcatenator. Otherwise you'd might end up with odiaCamb-
So I've added an extra sorting attribute, that's being filled with a conditional value: if Name ends in a - it's the first part, so it gets a 1, the rest must be the second part so it gets a 2.
And the entire workflow:
Note that this may produce strange results if the name of the country contains hyphens
+2Hello @a_afiaax
I simulated here your case with the XLS Reader.
First: filter by attribute Name with symbol " - " - transformer StringSearcher.
Step 2: Create a attribute Order for each results from Output port Matched ( transformer StringSearcher ).
Step 3: Sroter by attribute Order and after Aggregator by attribute Point_ID.
Attached the Workspace Template - Workspace_String.fmwt
Thanks,
Danilo
Hello Danilo,
1st AttrributeMangaer includes 2 attributes i cant understand them !.
The final result looks nice but missed 2 records (ID:10Z and ID:14B):
Do you have any idea to keep all IDs?
Thank you.
+56So the general approach would be PointOnPointOverlayer to build the list, ListConcatenator to join the name values and then a StringReplacer to get the - out. But you need to make sure the list is sorted correctly, either by making sure the points enter the PointOnPointOverlayer in the correct order or by sorting it before the ListConcatenator. Otherwise you'd might end up with odiaCamb-
So I've added an extra sorting attribute, that's being filled with a conditional value: if Name ends in a - it's the first part, so it gets a 1, the rest must be the second part so it gets a 2.
And the entire workflow:
Note that this may produce strange results if the name of the country contains hyphens
PointOnPointOverlayer? Do you need to do it spatially if there is an ID for the key?
+2So the general approach would be PointOnPointOverlayer to build the list, ListConcatenator to join the name values and then a StringReplacer to get the - out. But you need to make sure the list is sorted correctly, either by making sure the points enter the PointOnPointOverlayer in the correct order or by sorting it before the ListConcatenator. Otherwise you'd might end up with odiaCamb-
So I've added an extra sorting attribute, that's being filled with a conditional value: if Name ends in a - it's the first part, so it gets a 1, the rest must be the second part so it gets a 2.
And the entire workflow:
Note that this may produce strange results if the name of the country contains hyphens
Hi @redgeographics , it end up with odiaCamb :
In addition to missing 2 record and 2 attributes (ID and Name):
+44I'd create the combined names by aggregating based on the Point_ID after ordering and then merge them back to get the output required
Combine_part_string.fmwt
+37Assuming the points are ordered so that the pairs are together (regardless of internal order), this can be done with a single AttributeCreator using adjacent feature handling and conditionals.
adjacentstring.fmwt
+37Assuming the points are ordered so that the pairs are together (regardless of internal order), this can be done with a single AttributeCreator using adjacent feature handling and conditionals.
adjacentstring.fmwt
If the assumption is incorrect, then a sorter on Point_ID prior to the attributeCreator would do the trick.
+60PointOnPointOverlayer? Do you need to do it spatially if there is an ID for the key?
Not really (but hey, I'm a spatial person)
+60Hi @redgeographics , it end up with odiaCamb :
In addition to missing 2 record and 2 attributes (ID and Name):
Did you set the AttributeManager and Sorter up like the example?
+2Assuming the points are ordered so that the pairs are together (regardless of internal order), this can be done with a single AttributeCreator using adjacent feature handling and conditionals.
adjacentstring.fmwt
Hi @jdh
Thank you for this solution. It takes little pit to understand conditions but it amazing.
