So far is seems that there's no obvious way to convert large (50+ billion rows) table to another Vertica table with an ARRAY[FLOAT] column using SELECT ... GROUP BY key ... query (which should create a row with an array of values for all 10,000+ rows with the same key for each distinct key) given that:
1) STRING_TO_ARRAY argument can't be longer than 65,000 bytes (i.e LONG VARCHAR isn't supported - for non-obvious reasons) while required limit in this case is ~500KB (that size is supported by LISTAGG but somehow not by STRING_TO_ARRAY)
2) ARRAY_CAT is limited to only 2 arguments (each being 1D array or NULL)
3) EXPORT TO PARQUET apparently can't export VARCHARs longer than 64KB
4) EXPORT TO PARQUET apparently can't export native 1D arrays - like ARRAY[FLOAT]
5) UDx can't handle complex types (including ARRAY)
6) saving output of a SELECT query using vsql and then using COPY ... CSV (which can parse sufficiently long strings and convert those to required ARRAYs) isn't feasible due to the size of output of that SELECT query (several TB).
Any suggestions on how to do it with Vertica would be appreciated.
Question
How can a native ARRAY[FLOAT] be created from values in a group (in Vertica 10.0.1)?
Sign up
Already have an account? Login
Welcome to the Rocket Forum!
Please log in or register:
Employee Login | Registration Member Login | RegistrationEnter your E-mail address. We'll send you an e-mail with instructions to reset your password.