CODEWITHSUNDEEP
kdb+q
jeonw
jeonw

Reputation: 133

kdb - issue with reading floating numbers

I am reading covariance data from flat files. For that reason, not being able to fully read the floating number results in covarince not satisfying positive semi definite requirements.

For instance, this is one of the input from raw text: “-0.581050672”— no, actually raw text is this: -5.801050672E-01

When I read this into kdb and cast with F, it results in -0.50810507. When I do this for all and check the covariance, unfortunately it does not satisfy PSD constraints. Other hack I have been doing is to add small noise in Identity matrix…

Apart from this hack, is there way to read above data into proper floating number up to 9th digit? I tried \P and .Q.f but these only seem to work in Display.

Thank you

Upvotes: 0

Views: 481

Answers (1)

jeonw
jeonw

Reputation: 133

Sorry, does not seem like a kdb issue. Was exporting these data into different software and floating points were lost during this process. Thanks for pointer.

Upvotes: 0

Related Questions