|
The technical forum of the NBN Trust for use by the Network. It includes discussions and announcements on the NBN Atlas, Recorder 6 and Indicia |
You are not logged in. Please login or register.
Forum → Recorder Troubleshooting and Bug Reports → Recommended Taxon Sort Order
Some species have a completely different sort order type
Philaenus spumarius comes out as 7400022399. there's quite a few invertebrates like this.
Meanwhile in my spreadshett some are showing as just very big numbers as a scientific formula such as "1E+26" which is for Dictyna arundinacea.
Why do these species not have the standard format sort order?
There is also a what question. The new recomended taxon order isn't in any sort of taxon order. I'm not sure whether to try running Mike's batch update from the How Do I section. Perhaps this needs to be in the features request section. Can we have a taxon sort order that really is a taxon order. The problem is that we've been extracting data assuming the recomended taxon order is clever and new and almost right when in fact it is a serious backwards step. For the data we've been working on for the last two weeks it's back to Excel Vlookup for us.
The new sort order is a text field. It always has been, but as it was all numeric it may have been treated as a numeric field somewhere within your processing, although I can't duplicate this effect, this would explain why it is being converted to a number. Treated as a text field it will give an order which is taxonomic, Kingdom, Phylum, Class, Order, Family, Genus, Species and is alphabetical within these. There are two problems with it. Firstly, for groups where sub orders are important in the taxonomy, the families are in alphabetical order and this causes problems (eg.lepidoptera, with butterflies and moths). Secondly, within families taxa are in alphabetical order which doesn't take account of the taxonomic order present in some lists.
The Batch update will put the order back to exactly what it was before, including the problems and anomalies.
To run the Batch Update copy it into the Batch Update folder of R6 (with R6 closed). Open R6 and it should appear in the Tools/Run Batch Updates under the Sort_Order menu. There is nothing else to do. For the moment you will need to run this Batch Update after each Dictionary upgrade. Running a rebuild of the Taxon Name Index will restore the new version.
I am working with the NHM to get amore refined sort order where possible, but this is not easy to do. You can get so far with most groups and then it get very difficult and time consuming.
The best we can do at the moment is to give users the choice of which version to use. The new version which is a taxonimic order, or the old version with the anomalies.
Hi Mike
I am treating it as a text field but reptiles come out after mammals and molluscs come out at the end. Meanwhile there's lots of anomalies. I mentioned a couple. Andricus quercuscalicis has a taxon sort order of 001L00000000000000000000000000 so also comes out at the end and there are a vareity of insects and especially spiders coming out before lichens.
Then there are those which seem to have the old short number code instead of the new version so these all end up at the beginning. I did an index rebuild but they still have the old version sort numbers. These, in my selection were mosses, Hemiptera and Neuroptera. The new version has more anomalies than the old version.
This isn't what I would expect and certainly not what is reflected in my version of the Dictionary. All entries in Index_taxon_Name should end up with the new aplha numeric version of the sort order. I will need to give some thought to why this is happening. I suggest you revert back to the old sort using the Batch Update, while we work out what is going on.
Keith mentions a few plants with the old format in his post here and the spreadsheet shows the reptile and molluscs out of order.
http://forums.nbn.org.uk/viewtopic.php?id=4772
Hi Mike
The issue Graham is describing also happens in my copy. I mentioned to you after the 1Q upgrade that I don't have the same number of rows in the Organism table as you; could this still be affecting other users (http://forums.nbn.org.uk/viewtopic.php?id=4510)
The reptiles and molluscs (and various marine invertebrates) coming after mammals is still part of the alphabetical sorting issue;
within the 'Taxon Phylum' Chordata (Vertebrates) the 'Taxon Class' order is Amphibia, Aves (Birds), Mammalia, Reptilia. Fish are mixed up amongst all of these because they all fall in different classes, e.g. Actinopterygii, Cephalaspidomorphi;
the Phylum Chordata is then followed by the the Phylum Mollusca.
I have queried a large well-recorded site and ordered it by the current 1S Recommended Taxon Sort Order, listing Kingdom,
Phylum, Class and Order to demonstrate this.
I'm pleased to see that taxonomic order is getting a bit more of a profile at the moment at various places on the forum. I see you have sorted out the plant order in 1S; I was too busy to follow it up with you at the time, so thanks for that. I'm sure not everyone will be happy with it and it still needs refinement at the species level, but at least it is now following Stace/APG at the Genus level. Thanks again.
Alison
It does sound like the Orgnanism table may not have fully installed on all systems. There were lots of problems with the size of it intially which is why it had to be boken down into so many different upgrades. There should be 108337 rows in it after dictionary upgrade 000001S. If this isn't the case we need to look at how best to fix this. Thank you Graham for the file. The sort orders certainly shouldn't look like this and need investigating.
Mike, we have 480 records in the INDEX_TAXON_NAME table with SORT_ORDERs which are not 30 characters in length. 403 of these are taxa I've added myself (might they have SORT_ORDERS because I've used your Manage User Added Taxa addin) ?
In case its of any use, I've attached the 77 rows from the INDEX_TAXON_NAME with SORT_ORDERs which are not 30 characters in length (i.e. exlcuding the usesr added taxa).
(we will probably take you up on your offer Mike, but this will need to be planned carefully our end first!)
Thank you very much. The list is very helpful . Investigating thes shows a different reason for 73 of these, which I should be able to fix without having your database. I suspect this may affect all users except those installing from 6.17.
There are 4 taxon list item keys in there JNCC004500000001,2,3,C which don't seem like they came from the NHM so there may be another explanation for these.
Whoops, ignore JNCC004500000001,2,3,C - they are a user added taxa from a Recorder 3 transfer many moons ago!
In the light of the information provided by Charlie I don't think we should assume that problems exist until dictionaries are up to 0000001U. If at that point Dictionaries have any Recommended Taxon Sort orders which are not 30 characters in length, then action will be necessary to bring these dictionaries into line. Let me know if this is the case and I will provide details of what will be needed to resolve the problems. The small upgrade 1T,1U will be available on the Recorder6 site shortly. This will fix the 73 problems identified by Charlie and which will affect most systems.
On the generally question of sort orders, I have files available which will fix the issues at the higher level in the hierarchy and also files which implement sort orders at species for the most many groups. I will need to have these check by Chris Raper, before implementing them.
Hi Mike
Following the upgrade to 1U, I have 215 rows left with the incorrect format of sort order. I have attached a spreadsheet of those rows, but no doubt we will need to be taking you up on your offer. There are also a further 50 rows where more than one species have the same incorrect sort order.
Alison
The best way to identify an fix these will be to have a copy of your database. There will be no need to stop using recorder and I will provide a standard Dictionary upgrade for you to run.
Forum → Recorder Troubleshooting and Bug Reports → Recommended Taxon Sort Order
Powered by PunBB 1.4.6, supported by Informer Technologies, Inc.