1083 medical doctors
(Arno Müller vol. 5)

This page deals with the file 5muller_medics.txt downloaded from newalchemypress.com.
This file contains records already present in Cura files A2 and E1, and brings 224 new birth dates.
It was used to :
  • Fix column GNR and build a modified version : 1083MED.csv.
  • correct names and birth days in files A2 and E1
More work can be done on this file :
  • Check legal time restoration in A2.
  • Better place names in A2 (which would permit better matching to geonames).

Generalities

This group was built by Arno Müller and Suitbert Ertel in 1994.
Data were published by Arno Müller in "1083 members of the French Académie de Médecine", Astro-Forschungs-Daten 5, Waldmohr, 1994, 92 pp.
Data available in an electronic format on newalchemypress.com web site, through file 5muller_medics.txt.

Newalchemypress says that this file concerns one of the two independant replications of a Gauquelin planetary effect by Arno Müller. So this file has a particular historical interest.

An other interest of this file is the good quality of its data : full copy of birth certificate name with accents (not always), separation of family and given names, legal hour, timezone offset indications.

Integration to g5

This file is imported in g5 database but following doc has not been updated yet.
The raw file is data/raw/newalchemypress.com/05-muller-medics/5a_muller-medics-utf8.txt

Full execution

The full set of transformations can be executed with the command :
php run-g5.php newalch muller1083 all
This is equivalent to :
php run-g5.php newalch muller1083 raw2tmp
php run-g5.php newalch muller1083 tweak2tmp
php run-g5.php newalch muller1083 fixGnr update

Generate csv

The first step is to generate the file 5-newalch-csv/1083MED.csv :
php run-g5.php newalch muller1083 raw2tmp
This step includes minor corrections in given names (some accents are missing).

tweak2tmp

This step applies corrections from file data/3-edited/newalch-tweaked/1083MED.yml, see page tweak2tmp.
Used to fix errors in the names of noble persons (see below).

Fix GNR

Looking at the file with
php run-g5.php newalch muller1083 look
(see below) showed abnormal differences for some names. This led to identify errors on GNR :
Column GNR is truncated, so GNRs superior to 99 are possibly erroneous.
Step fixGnr showed that 87 records have a wrong GNR.
The command
php run-g5.php newalch muller1083 fixGnr report
lists the GNRs to fix with the corresponding values of dates and names in A2 and E1.
This step is necessary, to manually check that the fix does not introduce wrong associations (this is not the case).
Details
For each GNR > 99, the code builds a list af possible matches in Cura files (for example, A code 103 in Müller's file could correspond to 103, 1030, 1031 ... 1039.
To choose the right match, the code uses field DATE (in fact, it uses only the day).
But cura files contain errors in DATE field.
The improbable situation could have occured : among the 11 candidates, the good one contains an error in DATE, and an other candidate has Müller's date.
In this case, the code would have generated a mistake. So report with human check is necessary.
To correct GNR newalch-csv/1083MED.csw, run
php run-g5.php newalch muller1083 fixGnr update
It modifies 87 records.

Fix Cura

These steps are not part of Müller 1083MED restoration process, but are included in the processes of A2 and E1.

Fix Cura names

Step fixGnr must have been executed before this step.

The comparison between names in Müller's file and Cura (see below) shows that Müller's names are better than Cura's.
Step fixCura copies Müller's names in corresponding Cura records.
php run-g5.php newalch muller1083 fixCura
PARAMETER MISSING - This function needs 3 parameters :
  Param 1 can be : 'A2', 'E1'
  Param 2 can be : 'names', 'days'
  Param 3 can be : 'report', 'update'
To restore the names :
php run-g5.php newalch muller1083 fixCura A2 names update
739 records modified in data/5-tmp/cura-csv/A2.csv
12 unknown names restored in A2
php run-g5.php newalch muller1083 fixCura E1 names update
82 records modified in data/5-tmp/cura-csv/E1.csv
0 unknown names restored in E1
This permits to restore (only) 12 unknown names from A2.

Fix Cura days

Step fixGnr must have been executed before this step.

The comparison between birth days in Müller's file and Cura A2 and E1 (see below) shows that among the 859 common records, 39 have different birth days (4.54 %).

The only way to find out which day is correct is to check on birth certificates. Other sources are available, like Wikipedia or Comité des travaux historiques et scientifiques, but these sources may contain errors.
French registries are available online, so checking is possible, but this is a long process.
Very few records were really checked, and all of them showed that the error was coming from Cura file.
Müller NR 107 Blondot Nicolas 1880-02-04
Cura A2-80 Blondot Nicolas 1880-02-05
Checking online civil registry seems to indicate that the error comes from file A2.
Not a firm conclusion because the birth certificate is hard to understand.
But the declaration date (1808-02-05 16:00) correponds to the birth date in A2.
The declaration cannot be simultaneous with birth date, so A2 date is erroneous.
To check : Archives des Vosges, commune de Charmes, registre 4E92/7, p 14 / 22.
Birth certificate permits to check that Müller's hour is correct (16:00 - quatre heures du soir).

Müller NR 342 Fabre Jean 1864-06-08
Cura A2-255 Fabre Jean 1874-06-08
Checking online civil registry permits to conclude that the error comes from file A2.
Jean Fabre is born in 1864, in Lyon, 4th arrondissement.
See Table décennale and birth certificate : Lyon 4e Naissances - 2E10360 - p 56 / 129.
Birth certificate also permits to check that Müller's hour is correct (04:00).

Müller NR 484 Heuyer Georges Jean-baptiste 1884-01-30
Cura A2-1051 Heuyer Georges 1884-07-30
Checking online civil registry permits to conclude that the error comes from file A2.
Heuyer Georges Jean-baptiste is born 1884-01-30.
See birth certificate on registry 8 Mi 4969 (NMD 1883-1892) , Pacy-sur-Eure, p 55 / 585.
Birth certificate also permits to check that Müller's hour is correct (20:30).

Müller NR 743 Morice André Marie Gustave Emile Etienne 1890-07-31
Cura E1-1520 MORICE André 1891-07-31
Checking online civil registry permits to conclude that the error comes from file E1.
Morice André Marie Gustave Emile Etienne is born 1890-07-31.
See birth certificate on registry Etat-civil de Caen, Naissances 1890, p 141 / 251.
Birth certificate also permits to check that Müller's hour is correct (15:00).

Müller NR 914 Rocher Henri Gaston Louis 1876-05-28
Cura E1-1806 ROCHER Louis 1876-05-27
Checking online civil registry permits to conclude that the error comes from file E1.
Morice André Marie Gustave Emile Etienne is born 1876-05-28.
See birth certificate on registry 4 E 1548, p 99 / 274.
Birth certificate also permits to check that Müller's hour is correct (16:00).
The supposition was made that Müller's days are globally more exact than Cura's, which lead to inject Müller's days in A2 and E1.
This may lead to introduce errors in Cura data.

Note : birth hours were not modified because it deals with the problem of timezone offset, handled in an other step.

Other note : extrapolating the error rate of 4.54 % to all Cura data used in g5 program (21 249) leads to 964 erroneous dates in Cura files. This does not permit to suppose that this error rate was the same in Gauquelin's original data, because errors may have been introduced in the editing process of cura.free.fr.

Execution for A2 :
php run-g5.php newalch muller1083 fixCura A2 days update
33 records modified in data/5-tmp/cura-csv/A2.csv
for E1 :
php run-g5.php newalch muller1083 fixCura E1 days update
2 records modified in data/5-tmp/cura-csv/E1.csv

Fix Paris

The file contains 110 persons born in Paris, which need to be checked (see Newalchemypress indication below, sections SAMPLE and GAUQ_NUR).

Fix arrondissement

All people in Müller's file are noted Paris 1er arrondissement, which is not correct. There is a way to find the arrondissement : check in Cura file A2 and find the corresponding person, using birth date ; the arrondissement is noted in A2.
Once the arrondissement is found, it's possible to check in the online registries.

Check dates and times

Checking is possible using online archives, but there is one important restriction : part of the Paris archives were destroyed in 1871 (events known as "La commune de Paris", some acts and their backups were burnt), and acts prior to 1860 were lost. Partial reconstitutions of the acts could be done. Online data prior to 1860 contain birth day but no time. There may exist other reconstitutions containing birth times, but I don't know where (maybe present on sites demanding registration and / or pay, like familysearch.org or filae.com).

So I don't know where Gauquelin found the birth times of persons born in Paris before 1860.
php run-g5.php newalch muller1083 look paris
Shows that there are 70 persons born before 1860 ; only possible to check 110 - 70 = 40 persons.
7 have been checked so far, one error was found :
Balthazard Victor 1872-01-01 09:00
NR 47
A2-32
Paris 11
ERROR IN BIRTH TIME : 21:00 ("neuf heures du soir"), not 09:00

TODO

  • Include these corrections in data/3-edited/newalch-tweaked/1083MED.yml to be processed by step tweak2tmp.
  • Check the remaining records.

Fix nobilities

Listing names shows small errors in columns FNAME and GNAME of noble persons.
  • Sometimes the nobility is included in the given name (it should be part of the family name).
  • The nobility is uppercased (it should be lowercased).

To list the noble persons of the file.
php run-g5.php newalch muller1083 look nobilities simple
61 de Barthez| Antoine Charles Ernest
152 de Brun| Hippolyte Marie Antoine
(...)
1044 de Vernejoul| Robert
16 records concern noble persons.
The command :
php run-g5.php newalch muller1083 look nobilities yaml
prints the same list in a YAML format.

This was used to copy the output in file data/3-edited/newalch-tweaked/1083MED.yml.
Nobilities are then all fixed by step tweak2tmp.
The rest of this page concerns preparatory code
and is not part of any data transformation process.

Looking at the file

With class src/commands/newalch/muller1083/look.php :
php run-g5.php newalch muller1083 look
PARAMETER MISSING
Possible values for parameter : curadates, curanames, fields, gnr, nobilities, paris, sample
    

Fields

php run-g5.php newalch muller1083 look fields
Partial comprehension so far.
FieldMeaning
NRMüller id, from 1 to 1083
SAMPLEOrigin of the record
GNRGauquelin NUM in A2 or E1
CODE
NAMEFamily and given name
GEBDATUMBirth day
JAHRBirth year
GEBZEIT
GEBORT
LAENGE
BREITE
MODE
KORR
ELECTDATDate of election in Académie de médecine
ELECTAGEAge of election
STBDATUM
SONNE
MOND
VENUS
MARS
JUPITER
SATURN
SO_
MO_
VE_
MA_
JU_
SA_
PHAS_
AUFAB
NIENMO
NIENVE
NIENMA

Column SAMPLE

This column seems to indicate the origin of the record.
There is a one to one mapping between SAMPLE and CODE.
php run-g5.php newalch muller1083 look sample
to generate a html table :
php run-g5.php newalch muller1083 look sample table
SAMPLECODENbGNR ?
MUER_NUR1224N
MUERGAUQ-d279Y
MUERGAUQ3612Y
GAUQ_NUR4168Y
Indications from newalchemypress.com :
Only for those in Michel's original group born in Paris (168) was Müller unable to re-check their birth data (...) This new total of Academie members had 224 not in MG's original sample.

GAUQ_NUR

Newalchemypress indication about Paris could designate GAUQ_NUR, but it doesn't match :
GAUQ_NUR total168
GAUQ_NUR not born in Paris59
GAUQ_NUR born in Paris109
Total born in Paris110 (1 marked MUER_NUR)

MUER_NUR

All MUER_NUR records have GNR empty, unknown in Gauquelin data => corresponds to Newalchemypress indication.

MUERGAUQ

?

Column GNR

Seems to mean "Gauquelin Number Record".
Values starting by SA2 correspond to Cura file A2 (SA may mean "Serie A") ; for example SA232 corresponds to record NUM 32 in file A2.
Values starting by ND1 correspond to Cura file E1 (ND1 may mean "New Data 1") ; for example ND1204 corresponds to record NUM 204 in file E1.

This command counts the records associated to Gauquelin files :
php run-g5.php newalch muller1083 look gnr
EMPTY 	: 224
A2 	: 765
E1 	: 94
A2 + E1 : 859
Total 	: 1083

Comparing with Cura names

The command :
php run-g5.php newalch muller1083 look curanames
prints side by side family and given names of records present both in Müller's file and file A2.

This must be exectuted after having executed step fixGnr to see all names.
This shows that Müller's names are better. It also shows that spelling of name contain mistakes for noble persons (see below).

Comparing with Cura dates

The command
php run-g5.php newalch muller1083 look curadates
displays the differences between Müller's dates and A2 and E1.
Müller NR 31   1904-04-06	| Aubry	| Maurice Charles Louis
E1    NUM 45   1899-04-06	| AUBRY Maurice Charles	| 

Müller NR 107  1808-02-04	| Blondlot	| Nicolas
A2    NUM 80   1808-02-05	| Blondlot	| Nicolas

(... displays 39 differences ...)

Compare dates Müller / A2 E1
        | Equal         | Different  | Total
--------------------------------------------
A2      | 730 (95.42 %) | 35 (4.58 %)| 765
E1      | 90  (95.74 %) | 4  (4.26 %)| 94
A2 + E1 | 820 (95.46 %) | 39 (4.54 %)| 859
Note : these results were obtained from Cura files generated by raw2tmp. If step tweak2tmp is applied on A2 or E1 before looking at the differences, these results are different, because tweak2tmp include corrections on the concerned records.