Hashed file size based on its columns datatype/length

[datability_user]

Hi All,

Is there any difference in the file size of a hashed file, based on its columns datatype/length structure,
i.e,
HF1:
col1 varchar(100)/char(100),
col2 decimal(100)/integer(100)/numeric(100)

HF2:
col1 varchar(10)/char(10),
col2 decimal(10)/integer(10)/numeric(10)

HF3:
col1 varchar(1)/char(1),
col2 decimal(1)/integer(1)/numeric(1)

if the data from the source coming as 1 digit (alphanumeric)

thanks in advance

[ArndW]

Hashed files only have one internal datatype, namely a string. Length specifications are informational only, thus all 3 hashed files you mentioned will be identical.

CHAR fields will be padded with blanks in DS and written/read with blanks.

[datability_user]

Thanks ArndW very much
that means i do not have to care about defining data types and length for my hashed files, anything will be acceptable am i right?

[ArndW]

You should always use realistic numbers, particularly for non-strings; but in Server you can put "Hello There" into a VarChar(3) field and it would still be stored correctly.

[datability_user]

thanks again ArndW,

Why numbers?
if i stored, say 999 in integer(1), is it going to be stored correctly?

[chulett]

Yes.

[ArndW]

Even though DS Server does not restrict you from misusing job column metadata it is always best to act as if it would and use column definitions that approximate the data; that way if you were to write to a database you can avoid error messages upon insert and update.

[chulett]

Exactly. At some point it will be enforced, after all.

[datability_user]

Again thanks,
i will flag this as resolved