blob: 3905bdbaf2600d0a93b6a8708c62bd55c10f676b (
plain) (
blame)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
|
This note describes some specific characteristics and limitations
of STSDAS tables.
1. Six data types are supported:
single-precision real
double-precision real
integer (same number of bytes as real)
short integer (int*2)
boolean (Fortran logical; same size as real)
text (character strings) in multiples of four characters up to
160 characters
Type conversion is done implicitly if the data type of the column differs
from that of the buffer. Boolean true and false are converted to YES and NO
respectively if they are read as integer values.
2. Header parameters:
The same data types are supported for header (user) parameters as for
table data, except for short integers, and text strings are limited to 70
characters. Type conversion is done for numeric and boolean parameters,
but text parameters may only be gotten as text. The I/O routines for header
parameters get and put single parameters, not arrays. A user program may
construct array elements by appending numbers to a root portion of a keyword,
and the array elements may be gotten or put one at a time.
The table must exist in order to get or put header parameters.
3. Undefined elements are flagged by using special values. There is at
present no undefined value for a boolean element; the default value is false
(or no). A null character string is regarded as undefined. Trailing blanks
are truncated by the Fortran-callable routines (ut...) but not by the
SPP-callable (tb...) routines.
4. The format for printing table values supports the more common Fortran
formats with a few differences and some additional options:
h: H:M:S.d
m: H:M.d (or M:S.d)
i-: left justified (but SPP uses d instead of i)
i0: fill the field with zeros on the left; Fortran also allows filling
a portion of the field (e.g. I4.2)
o: octal
x: hexadecimal
b: "yes" or "no" instead of "T" or "F"
5. The header parameters and column descriptors are stored in the table
together with the data, so if more parameters are written or more columns
defined than there is space in the table, the table will be rewritten to
allocate more room. If the table is large, this can be a problem either
due to the time involved or because of disk space limitations. When a new
table is created it is possible to specify the amount of space to be
allocated for header parameters and for column descriptors.
A table may be row-ordered (which is the default) or column-ordered.
If the table is row-ordered then rows may be written at the end of the table
without ever specifying the maximum number of rows. Adding new columns after
creating the table, however, requires rewriting the table unless space was
allocated for new columns.
There is no intrinsic limit on the number of columns or rows.
|
