Skip to content

Latest commit

 

History

History
36 lines (23 loc) · 1.36 KB

cookbook.md

File metadata and controls

36 lines (23 loc) · 1.36 KB
Raw

Cookbook

Get the best yummy yummy generated scripts ....

Not using all fields from the CSV file

Create a hint with the wanted column count, but with a null label name

 List<Hint> hints = new ArrayList<>();
 hints.add( new Hint(null,2) );
 hints.add( new Hint("Country",2) );
 hints.add( new Hint("Planet",2) );
 skel.setHints( hints );

The dreaded SIREN file !

The list of all French companies is published as open data. It has 130+ columns. It is what made me create this tool. The command I propose is

CALL wadael.csvskelgen("/home/jerome/Tools/neo/neo4j-community-3.3.1/import/siren78.csv",";",4,"Etablissement:16,Localisation:6,Region:2,Etablissement:16,NatureEI:2,Activite:3,Effectif:4,Etablissement:11,Entreprise:2,Responsable:4,Entreprise:4,NatureJuridique:2,Etablissement:10,Categorie:1,Etablissement:12")

Then, you should tweak the output to group all the SET commands per label. As each group of column is treated as if if was a new label, some search/replace is needed.

Troubleshooting

This procedure is intended for use by developers. Consequently, it is not as idiocy-proof as a mainstream product.

If the result is an exception, check your parameters.

  • is it the good file ? path ?
  • is the separator correct ?
  • is the sum of the column counts given as hints <= to the number of columns of the file ?