News on Database errors project (2 October 2016)

Question

Analysis was done on data from October 2nd 2016.

Comments

Y'all are the "best" !!!  Just went to the list for the first time in quite a bit and did what I could.  

Thanks for doing this so faithfully for us.

- Betsy Stinson (Blevins-707)

Answer 1

Changes to search section

On http://wikitree.sdms.si/default.htm I updated Search group. Search was extended to all name fields, so you can now find also all Earl of Erroll texts.

There are added controls to include relatives on result page, Results are shown on multiple pages, and you can control number of items on page and sort order. There are also hyperlinks to the profiles found.

Comments

Great work, Aleš! Thanks for updating/tweaking your pages.

Answer 2

 

New error 811 Uncleaned profile after merge

I added new error, that lists profiles, that weren't cleaned after merge. For now I check only double sources header. 800 errors will be all biography errors.

Errors	Total	0000-0000	0001-1499	1500-1699	1700-1799	1800-1899	1900-1999	2000-Now	Open	New
811 Uncleaned profile after merge	206034	13702	7650	31731	45899	80080	26841	131	153320	206034

Comments

this is great!

can you also check for missing, mismatched or duplicated <ref> tags?

---

Ouch!  Aleš you are generating new types of issues faster than we can clear up the existing ones!  Perhaps you can help prioritize these 200,000 uncleaned-merge profiles by saying how many have 3 or more sources, rather than a mere two sources?

---

Paul,

I can, but there is really no point in that.I will add number od sources in report.

180000 have 2 sources, 18500 have 3 source sections, 4180 have 4 sources, 1250 have 5 sources and 1118 have more then 5,

And these are the winners:

Wikitree ID	Count
UNKNOWN-185515	77
Jordaan-290	59
Congn-2	42
Neville-335	32
Maine-10	32
Ufkes-53	29
Hill-4072	27
Merovingian-61	26
UNKNOWN-180463	22

---

Missing is very common (all unsourced profiles), So I dont think to add those at the moment.

Mismatched could be checked. 

Is new line allowed between <ref></ref> tags?

Duplicated If you mean double, I can. <ref><ref></ref></ref>

Also missing <references/> if <ref> are present.

---

Draft of help page created See Space:DBE_811

---

Paul, I added number of unmergerd profile in second column of the report. I also changed sort order, so they are listed first.

---

yeah, by missing, I really meant mismatched (opening without closing, etc.) :)

yes, new lines are allowed in between

each opening tag should only have one closing tag, and I don't think nesting is (or should be) allowed.

---

Awesome work, Aleš!

---

Thanks for the added column, that is helpful.

BTW, went through the Winner list of multiple sources above.  Messaged profile managers or cleaned up profiles as possible.  Following are pre-1500 so need some other person to assist:  Congn-2, Neville-335, Maine-10, Merovingian-61

Answer 3

Updated error 7X4 Wrong word in name

I added whole group of words to this error. First, second, ... They should be corrected to 1st , 2nd, ... form. I didn't include First wife, Second child, ... I added only those, that also have word Of in name (First Earl of something). Error was discussed here  .

	Total	Open	New
714 Wrong word in Prefix	945	469	1
724 Wrong word in First Name	7708	5341	57
734 Wrong word in Preferred Name	1790	164	57
744 Wrong word in Middle Name	964	829	20
754 Wrong word in Nicknames	1039	896	129
764 Wrong word in Suffix	1330	682	2
774 Wrong word in Last Name at Birth	14	14	14
784 Wrong word in Current Last Name	301	291	301
794 Wrong word in Last Name Other	72	68	72

Comments

Draft of help pages

DBE 714 714 Wrong word in Prefix DBE 724 724 Wrong word in First Name DBE 734 734 Wrong word in Preferred Name DBE 744 744 Wrong word in Middle Name DBE 754 754 Wrong word in Nicknames

DBE 764 764 Wrong word in Suffix DBE 774 774 Wrong word in Last Name at Birth DBE 784 784 Wrong word in Current Last Name DBE 794 794 Wrong word in Last Name Other

 

Answer 4

New errors 801 Big profile. 802 Empty profile and 803 Almost empty profile

I added these errors to list profiles just by sizes. 801 are profiles bigger than 200000 letters, 802 are empty profiles, and 803 are very small profiles, less than 50 letters.

Errors	Total	0000-0000	0001-1499	1500-1699	1700-1799	1800-1899	1900-1999	2000-Now	Open	New
801 Big profile	324	5		7	13	153	146		241	324
802 Empty profile	86485	35687	1062	1833	6758	29831	11242	72	43030	86485
803 Almost empty profile	16819	3560	657	821	2057	6592	3118	14	11671	16819

Comments

These are great, Ales. We can use them for various challenge ideas. Thank you for all the time you put into generating these each week!

---

Draft of help pages done. Empty profile is for me an unsourced profile that must be sourced. Seems easier that we mark them {{Unsourced}} ==> we increase the unsourced profile Category in WikiTree with +103304 profiles ==>  Unsourced Profiles ‎(346 149 members)

• DBE 801 801 - Big profile
• DBE 802 802 - Empty profile
• DBE 803 803 - Almost empty profiles

 

---

Error 1014 - Database Error program has errors

Error 802 

• Err_20161002/802_60_0.htm
• Err_20161002/802_2000-Now_0.htm
• Err_20161002/802_1900-1999_0.htm
• Err_20161002/802_New_0.htm

---

Fixed

---

Thanks

@Aleš we should have a WikiTree #Quick Statement to add template  {{Unsourced}} to those profiles as today that is the standard for marking unsourced profiles inside WikiTree...


Found this one -Data_Entry_Mistake--1 you never stop to get surprised.... at least born in London seems like a good start.... looks like someone tries to delete a profile...

---

This is such a beautiful feature that you've built Ales, have been going through it for some hours now ... finding and fixing some that I recognize as my own or other South African profiles (only the Twentieth century profiles I leave as it is, figuring the managers should take responsibility for their profiles).

One question though - I have found a few 18th century profiles and after {{DateGuessing2}} and then comparing again on [variants] of surnames, found matches and proposed mergers. I then "hide error" ... I understood in a previous conversation with you that an error (if there is no possibility of making it into a "false" error) that has been hidden, will remain that way [edit: I understood it will show up again as an error after 1 month; there is no undoing it as an error mention]. I have added to many profiles {{Unsourced| South Africa}}{{DateGuess2}} and proposed some merges for some [but not for all] but what now if after a month, it shows up again as an error (with no option to make it "false") but is has some body to the bio, even validated?

---

Philip, which error are you reffering to?

---

I forgot to put in the link. Specifically I'm referring to this one: http://www.softdata.si/osebe_staro/ales/wikitree/Err_20161002/802_0000-0000_8.htm (it also contains profiles which by nature of the absence of dates on some profiles eventually fall into for arguments' sake the 18th century) though I have also been busy here: http://www.softdata.si/osebe_staro/ales/wikitree/Err_20161002/802_1700-1799_0.htm

 

I still have to go through the database(s) with almost empty profiles.

---

I don't think 802 and 803 deserves False error. Some bio and sources should be added. For 803 I intend to raise the limit. It is 50 letters now and I think it should be something like 500 letters minimum. I can also exclude unsourced profiles from 803. We will review the minimum size for a profile to be OK in following months. Also for error 801 I think we should lover the limit to 100000 letters or even less.

Back to your question, profile will simply move from 802 to 803 error if some text is added.

---

I changed Space:DBE_802 Space:DBE_803

no false error possible
{{Db_errors_G2G|803|N}}