Analyzing Dynamic Hash File

[asitagrawal]

Hi,

I am trying to use ANALYZE.FILE D_HashTest

But I am getting the following error:
Illegal option found on command line: D_HashTest

Please note that D_HashTest is a dynamic hased file.

Plz suggest.
Thx

[chulett]

No, it's not... but I'll wager HashTest is.

[asitagrawal]

Hi,

I am trying to use ANALYZE.FILE D_HashTest

But I am getting the following error:
Illegal option found on command line: D_HashTest

Please note that D_HashTest is a dynamic hased file.

Plz suggest.
Thx

Another Query:

Can someone please explain how the number of bytes calculated from the field length? Please chk the below URL...

http://asit.agrawal.googlepages.com/Query3.jpg

Thx

[ray.wurlod]

That is output from the HASH.HELP.DETAIL command or something similar. The record length is not calculated from the field sizes; it is actually measured from the data stored in the file.

What you have highlighted in the picture is merely the key columns. Chances are that there are non-key columns as well. These will contribute to the reported record size.

This utility does not take into account physical storage overheads, which may include any or all of:

• record header (12,20,24 or 40 bytes)
  
  segment mark between key and data
  
  padding to whole word boundary

However I believe it does take into account the dynamic array delimiter characters within the data record itself.

Hashed File Calculator does attempt to take the storage overheads into account, except for the extra for oversized records.

By the way, what you analyzed is NOT a dynamic hashed file; it's a static hashed file (Type 3).

[asitagrawal]

That is output from the HASH.HELP.DETAIL command or something similar. The record length is not calculated from the field sizes; it is actually measured from the data stored in the file.

What you ...

Hi Ray,

Please trhow some light on relation between the data stored and the bytes, for a particular record (row).
I am tryin to use HFC utility, where I have to feed the Record Size and there it says, Record Size can be estimated from the Fields' Display Size and I have used the field display size same as the field length.

Regds
Asit

[ray.wurlod]

You have to count all fields, not just the key columns.
There is a single byte delimiter between each field. Count those too.
That is the figure to present to HFC - it does the rest.

[asitagrawal]

You have to count all fields, not just the key columns.
There is a single byte delimiter between each field. Count those too.
That is the figure to present to HFC - it does the rest. ...

Hi Ray,
Can u plz confirm/correct the following:

Char - 1 Byte
Integer - 4 Bytes
SmallInt - 2 Bytes
Date - ?
Decimal - ?

Thx

[ray.wurlod]

You are not correct.
All data are stored as strings.
Char(N) = N bytes without NLS, up to 4N bytes with NLS
Integer = number of digits (therefore up to 10 characters, plus see above)
and so on.
That is why we advise to use the display width to estimate.