Negative number in the contengy table count

[Nolife999]

The files are quite large so I will have to give them the test case by another way than github.

Before that, I was thinking it is maybe just a usage of blocking variable problem. Could you tell me if the files have to be ordered on the blocking key before processing with RELAIS ?

[luvalent]

No it is not required

[luvalent]

what do you mean with 'Negative number in the contengy table count'?

[Nolife999]

I will give you an example Luca; juste big files

[Nolife999]

SIRET_APP	FREQUENCY
0	-1905011591
1	232457

1 REDUCTION METHOD {"REDUCTION-METHOD":"BlockingVariables","BLOCKING":{"BLOCKING A":["BDD_SIR_7_TYPEQU"],"BLOCKING B":["SRC_SIR_7_TYPEQU"]},"SORTED NEIGHBORHOOD":{"SORTING A":[],"SORTING B":[]},"SIMHASH":{"SHINGLING A":[],"SHINGLING B":[],"HDTHRESHOLD":"30","ROTATIONS":" 4"}}
2 MATCHING VARIABLES
MatchingVariable: SIRET_APP MatchingVariableA: BDD_SIR_7_SIRET MatchingVariableB: SRC_SIR_7_SIRET Method: Equality

[luvalent]

Thank you Manuel! This seems a bug in creating the contingency table because is not possible a negative value a this level... Can you tell me how many records are in the two datasets? and how many values can take the variable BDD_SIR_7_TYPEQU? Luca

…

----- Messaggio originale ----- Da: "Manuel Soulier" <notifications@github.com> A: "mecdcme/is2" <is2@noreply.github.com> Cc: "Luca Valentino" <luvalent@istat.it>, "Comment" <comment@noreply.github.com> Inviato: Venerdì, 5 marzo 2021 9:04:39 Oggetto: Re: [mecdcme/is2] Negative number in the contengy table count (#39) SIRET_APP | FREQUENCY -- | -- 0 | -1905011591 1 | 232457 1 REDUCTION METHOD {"REDUCTION-METHOD":"BlockingVariables","BLOCKING":{"BLOCKING A":["BDD_SIR_7_TYPEQU"],"BLOCKING B":["SRC_SIR_7_TYPEQU"]},"SORTED NEIGHBORHOOD":{"SORTING A":[],"SORTING B":[]},"SIMHASH":{"SHINGLING A":[],"SHINGLING B":[],"HDTHRESHOLD":"30","ROTATIONS":" 4"}} 2 MATCHING VARIABLES MatchingVariable: SIRET_APP MatchingVariableA: BDD_SIR_7_SIRET MatchingVariableB: SRC_SIR_7_SIRET Method: Equality

-- You are receiving this because you commented. Reply to this email directly or view it on GitHub: https://urlsand.esvalabs.com/?u=https%3A%2F%2Fgithub.com%2Fmecdcme%2Fis2%2Fissues%2F39%23issuecomment-791238491&e=3cfb7ead&h=2b1a0d55&f=y&p=y

[Nolife999]

Hi Luca, I don't forget you. Just lot of work these days but I could answer you beginning of the next week.
Manu

[luvalent]

ok manu, no problem :-)

[francescoamato]

Hi Manu,
probably, if you have a large dataset, the negative fequency is a overflow problem.
I changed the frequency type in Long.
Can you try...

Francesco