1 @node System File Format
2 @appendix System File Format
4 A system file encapsulates a set of cases and dictionary information
5 that describes how they may be interpreted. This chapter describes
6 the format of a system file.
8 System files use four data types: 8-bit characters, 32-bit integers,
10 and 64-bit floating points, called here @code{char}, @code{int32},
12 @code{flt64}, respectively. Data is not necessarily aligned on a word
13 or double-word boundary: the long variable name record (@pxref{Long
14 Variable Names Record}) and very long string records (@pxref{Very Long
15 String Record}) have arbitrary byte length and can therefore cause all
16 data coming after them in the file to be misaligned.
18 Integer data in system files may be big-endian or little-endian. A
19 reader may detect the endianness of a system file by examining
20 @code{layout_code} in the file header record
21 (@pxref{layout_code,,@code{layout_code}}).
23 Floating-point data in system files may nominally be in IEEE 754, IBM,
24 or VAX formats. A reader may detect the floating-point format in use
25 by examining @code{bias} in the file header record
26 (@pxref{bias,,@code{bias}}).
28 PSPP detects big-endian and little-endian integer formats in system
29 files and translates as necessary. PSPP also detects the
30 floating-point format in use, as well as the endianness of IEEE 754
31 floating-point numbers, and translates as needed. However, only IEEE
32 754 numbers with the same endianness as integer data in the same file
33 has actually been observed in system files, and it is likely that
34 other formats are obsolete or were never used.
36 The PSPP system-missing value is represented by the largest possible
37 negative number in the floating point format (@code{-DBL_MAX}). Two
38 other values are important for use as missing values: @code{HIGHEST},
39 represented by the largest possible positive number (@code{DBL_MAX}),
40 and @code{LOWEST}, represented by the second-largest negative number
41 (in IEEE 754 format, @code{0xffeffffffffffffe}).
43 System files are divided into records, each of which begins with a
44 4-byte record type, usually regarded as an @code{int32}.
46 The records must appear in the following order:
56 All pairs of value labels records and value label variables records,
60 Document record, if present.
63 Any records not explicitly included in this list, in any order.
66 Dictionary termination record.
72 Each type of record is described separately below.
75 * File Header Record::
77 * Value Labels Records::
79 * Machine Integer Info Record::
80 * Machine Floating-Point Info Record::
81 * Multiple Response Sets Records::
82 * Variable Display Parameter Record::
83 * Long Variable Names Record::
84 * Very Long String Record::
85 * Character Encoding Record::
86 * Long String Value Labels Record::
87 * Data File and Variable Attributes Records::
88 * Extended Number of Cases Record::
89 * Miscellaneous Informational Records::
90 * Dictionary Termination Record::
94 @node File Header Record
95 @section File Header Record
97 The file header is always the first record in the file. It has the
104 int32 nominal_case_size;
109 char creation_date[9];
110 char creation_time[8];
116 @item char rec_type[4];
117 Record type code, set to @samp{$FL2}.
119 @item char prod_name[60];
120 Product identification string. This always begins with the characters
121 @samp{@@(#) SPSS DATA FILE}. PSPP uses the remaining characters to
122 give its version and the operating system name; for example, @samp{GNU
123 pspp 0.1.4 - sparc-sun-solaris2.5.2}. The string is truncated if it
124 would be longer than 60 characters; otherwise it is padded on the right
128 @item int32 layout_code;
129 Normally set to 2, although a few system files have been spotted in
130 the wild with a value of 3 here. PSPP use this value to determine the
131 file's integer endianness (@pxref{System File Format}).
133 @item int32 nominal_case_size;
134 Number of data elements per case. This is the number of variables,
135 except that long string variables add extra data elements (one for every
136 8 characters after the first 8). However, string variables do not
137 contribute to this value beyond the first 255 bytes. Further, system
138 files written by some systems set this value to -1. In general, it is
139 unsafe for systems reading system files to rely upon this value.
141 @item int32 compressed;
142 Set to 1 if the data in the file is compressed, 0 otherwise.
144 @item int32 weight_index;
145 If one of the variables in the data set is used as a weighting
146 variable, set to the dictionary index of that variable, plus 1
147 (@pxref{Dictionary Index}). Otherwise, set to 0.
150 Set to the number of cases in the file if it is known, or -1 otherwise.
152 In the general case it is not possible to determine the number of cases
153 that will be output to a system file at the time that the header is
154 written. The way that this is dealt with is by writing the entire
155 system file, including the header, then seeking back to the beginning of
156 the file and writing just the @code{ncases} field. For files in which
157 this is not valid, the seek operation fails. In this case,
158 @code{ncases} remains -1.
162 Compression bias, ordinarily set to 100. Only integers between
163 @code{1 - bias} and @code{251 - bias} can be compressed.
165 By assuming that its value is 100, PSPP uses @code{bias} to determine
166 the file's floating-point format and endianness (@pxref{System File
167 Format}). If the compression bias is not 100, PSPP cannot auto-detect
168 the floating-point format and assumes that it is IEEE 754 format with
169 the same endianness as the system file's integers, which is correct
170 for all known system files.
172 @item char creation_date[9];
173 Date of creation of the system file, in @samp{dd mmm yy}
174 format, with the month as standard English abbreviations, using an
175 initial capital letter and following with lowercase. If the date is not
176 available then this field is arbitrarily set to @samp{01 Jan 70}.
178 @item char creation_time[8];
179 Time of creation of the system file, in @samp{hh:mm:ss}
180 format and using 24-hour time. If the time is not available then this
181 field is arbitrarily set to @samp{00:00:00}.
183 @item char file_label[64];
184 File label declared by the user, if any (@pxref{FILE LABEL,,,pspp,
185 PSPP Users Guide}). Padded on the right with spaces.
187 @item char padding[3];
188 Ignored padding bytes to make the structure a multiple of 32 bits in
189 length. Set to zeros.
192 @node Variable Record
193 @section Variable Record
195 There must be one variable record for each numeric variable and each
196 string variable with width 8 bytes or less. String variables wider
197 than 8 bytes have one variable record for each 8 bytes, rounding up.
198 The first variable record for a long string specifies the variable's
199 correct dictionary information. Subsequent variable records for a
200 long string are filled with dummy information: a type of -1, no
201 variable label or missing values, print and write formats that are
202 ignored, and an empty string as name. A few system files have been
203 encountered that include a variable label on dummy variable records,
204 so readers should take care to parse dummy variable records in the
205 same way as other variable records.
207 @anchor{Dictionary Index}
208 The @dfn{dictionary index} of a variable is its offset in the set of
209 variable records, including dummy variable records for long string
210 variables. The first variable record has a dictionary index of 0, the
211 second has a dictionary index of 1, and so on.
213 The system file format does not directly support string variables
214 wider than 255 bytes. Such very long string variables are represented
215 by a number of narrower string variables. @xref{Very Long String
216 Record}, for details.
222 int32 n_missing_values;
227 /* @r{Present only if @code{has_var_label} is 1.} */
231 /* @r{Present only if @code{n_missing_values} is nonzero}. */
232 flt64 missing_values[];
236 @item int32 rec_type;
237 Record type code. Always set to 2.
240 Variable type code. Set to 0 for a numeric variable. For a short
241 string variable or the first part of a long string variable, this is set
242 to the width of the string. For the second and subsequent parts of a
243 long string variable, set to -1, and the remaining fields in the
244 structure are ignored.
246 @item int32 has_var_label;
247 If this variable has a variable label, set to 1; otherwise, set to 0.
249 @item int32 n_missing_values;
250 If the variable has no missing values, set to 0. If the variable has
251 one, two, or three discrete missing values, set to 1, 2, or 3,
252 respectively. If the variable has a range for missing variables, set to
253 -2; if the variable has a range for missing variables plus a single
254 discrete value, set to -3.
257 Print format for this variable. See below.
260 Write format for this variable. See below.
263 Variable name. The variable name must begin with a capital letter or
264 the at-sign (@samp{@@}). Subsequent characters may also be digits, octothorpes
265 (@samp{#}), dollar signs (@samp{$}), underscores (@samp{_}), or full
266 stops (@samp{.}). The variable name is padded on the right with spaces.
268 @item int32 label_len;
269 This field is present only if @code{has_var_label} is set to 1. It is
270 set to the length, in characters, of the variable label. The
271 documented maximum length varies from 120 to 255 based on SPSS
272 version, but some files have been seen with longer labels. PSPP
273 accepts longer labels and truncates them to 255 bytes on input.
276 This field is present only if @code{has_var_label} is set to 1. It has
277 length @code{label_len}, rounded up to the nearest multiple of 32 bits.
278 The first @code{label_len} characters are the variable's variable label.
280 @item flt64 missing_values[];
281 This field is present only if @code{n_missing_values} is nonzero. It
282 has the same number of 8-byte elements as the absolute value of
283 @code{n_missing_values}. Each element is interpreted as a number for
284 numeric variables (with HIGHEST and LOWEST indicated as described in
285 the chapter introduction). For string variables of width less than 8
286 bytes, elements are right-padded with spaces; for string variables
287 wider than 8 bytes, only the first 8 bytes of each missing value are
288 specified, with the remainder implicitly all spaces.
290 For discrete missing values, each element represents one missing
291 value. When a range is present, the first element denotes the minimum
292 value in the range, and the second element denotes the maximum value
293 in the range. When a range plus a value are present, the third
294 element denotes the additional discrete missing value.
297 The @code{print} and @code{write} members of sysfile_variable are output
298 formats coded into @code{int32} types. The least-significant byte
299 of the @code{int32} represents the number of decimal places, and the
300 next two bytes in order of increasing significance represent field width
301 and format type, respectively. The most-significant byte is not
302 used and should be set to zero.
304 Format types are defined as follows:
307 @multitable {Value} {@code{DATETIME}}
393 @node Value Labels Records
394 @section Value Labels Records
396 The value label records documented in this section are used for
397 numeric and short string variables only. Long string variables may
398 have value labels, but their value labels are recorded using a
399 different record type (@pxref{Long String Value Labels Record}).
401 The value label record has the following format:
407 /* @r{Repeated @code{label_cnt} times}. */
414 @item int32 rec_type;
415 Record type. Always set to 3.
417 @item int32 label_count;
418 Number of value labels present in this record.
421 The remaining fields are repeated @code{count} times. Each
422 repetition specifies one value label.
426 A numeric value or a short string value padded as necessary to 8 bytes
427 in length. Its type and width cannot be determined until the
428 following value label variables record (see below) is read.
430 @item char label_len;
431 The label's length, in bytes. The documented maximum length varies
432 from 60 to 120 based on SPSS version. PSPP supports value labels up
436 @code{label_len} bytes of the actual label, followed by up to 7 bytes
437 of padding to bring @code{label} and @code{label_len} together to a
438 multiple of 8 bytes in length.
441 The value label record is always immediately followed by a value label
442 variables record with the following format:
451 @item int32 rec_type;
452 Record type. Always set to 4.
454 @item int32 var_count;
455 Number of variables that the associated value labels from the value
456 label record are to be applied.
459 A list of dictionary indexes of variables to which to apply the value
460 labels (@pxref{Dictionary Index}). There are @code{var_count}
463 String variables wider than 8 bytes may not be specified in this list.
466 @node Document Record
467 @section Document Record
469 The document record, if present, has the following format:
478 @item int32 rec_type;
479 Record type. Always set to 6.
482 Number of lines of documents present.
484 @item char lines[][80];
485 Document lines. The number of elements is defined by @code{n_lines}.
486 Lines shorter than 80 characters are padded on the right with spaces.
489 @node Machine Integer Info Record
490 @section Machine Integer Info Record
492 The integer info record, if present, has the following format:
504 int32 version_revision;
506 int32 floating_point_rep;
507 int32 compression_code;
509 int32 character_code;
513 @item int32 rec_type;
514 Record type. Always set to 7.
517 Record subtype. Always set to 3.
520 Size of each piece of data in the data part, in bytes. Always set to 4.
523 Number of pieces of data in the data part. Always set to 8.
525 @item int32 version_major;
526 PSPP major version number. In version @var{x}.@var{y}.@var{z}, this
529 @item int32 version_minor;
530 PSPP minor version number. In version @var{x}.@var{y}.@var{z}, this
533 @item int32 version_revision;
534 PSPP version revision number. In version @var{x}.@var{y}.@var{z},
537 @item int32 machine_code;
538 Machine code. PSPP always set this field to value to -1, but other
541 @item int32 floating_point_rep;
542 Floating point representation code. For IEEE 754 systems this is 1.
543 IBM 370 sets this to 2, and DEC VAX E to 3.
545 @item int32 compression_code;
546 Compression code. Always set to 1.
548 @item int32 endianness;
549 Machine endianness. 1 indicates big-endian, 2 indicates little-endian.
551 @item int32 character_code;
552 @anchor{character-code} Character code. The following values have
553 been actually observed in system files:
560 The @code{windows-1250} code page for Central European and Eastern
564 The @code{windows-1252} code page for Western European languages.
573 The following additional values are known to be defined:
586 Other Windows code page numbers are known to be generally valid.
588 Old versions of SPSS always wrote value 2 in this field, regardless of
589 the encoding in use. Newer versions also write the character encoding
590 as a string (see @ref{Character Encoding Record}).
593 @node Machine Floating-Point Info Record
594 @section Machine Floating-Point Info Record
596 The floating-point info record, if present, has the following format:
612 @item int32 rec_type;
613 Record type. Always set to 7.
616 Record subtype. Always set to 4.
619 Size of each piece of data in the data part, in bytes. Always set to 8.
622 Number of pieces of data in the data part. Always set to 3.
625 The system missing value.
628 The value used for HIGHEST in missing values.
631 The value used for LOWEST in missing values.
634 @node Multiple Response Sets Records
635 @section Multiple Response Sets Records
637 The system file format has two different types of records that
638 represent multiple response sets (@pxref{MRSETS,,,pspp, PSPP Users
639 Guide}). The first type of record describes multiple response sets
640 that can be understood by SPSS before version 14. The second type of
641 record, with a closely related format, is used for multiple dichotomy
642 sets that use the CATEGORYLABELS=COUNTEDVALUES feature added in
652 /* @r{Exactly @code{count} bytes of data.} */
657 @item int32 rec_type;
658 Record type. Always set to 7.
661 Record subtype. Set to 7 for records that describe multiple response
662 sets understood by SPSS before version 14, or to 19 for records that
663 describe dichotomy sets that use the CATEGORYLABELS=COUNTEDVALUES
664 feature added in version 14.
667 The size of each element in the @code{mrsets} member. Always set to 1.
670 The total number of bytes in @code{mrsets}.
673 A series of multiple response sets, each of which consists of the
678 The set's name (an identifier that begins with @samp{$}).
681 An equals sign (@samp{=}).
684 @samp{C} for a multiple category set, @samp{D} for a multiple
685 dichotomy set with CATEGORYLABELS=VARLABELS, or @samp{E} for a
686 multiple dichotomy set with CATEGORYLABELS=COUNTEDVALUES.
689 For a multiple dichotomy set with CATEGORYLABELS=COUNTEDVALUES, a
690 space, followed by a number expressed as decimal digits, followed by a
691 space. If LABELSOURCE=VARLABEL was specified on MRSETS, then the
692 number is 11; otherwise it is 1.@footnote{This part of the format may
693 not be fully understood, because only a single example of each
694 possibility has been examined.}
697 For either kind of multiple dichotomy set, the counted value, as a
698 positive integer count specified as decimal digits, followed by a
699 space, followed by as many string bytes as specified in the count. If
700 the set contains numeric variables, the string consists of the counted
701 integer value expressed as decimal digits. If the set contains string
702 variables, the string contains the counted string value. Either way,
703 the string may be padded on the right with spaces (older versions of
704 SPSS seem to always pad to a width of 8 bytes; newer versions don't).
710 The multiple response set's label, using the same format as for the
711 counted value for multiple dichotomy sets. A string of length 0 means
712 that the set does not have a label. A string of length 0 is also
713 written if LABELSOURCE=VARLABEL was specified.
719 The names of the variables in the set, each separated from the
720 previous by a single space.
723 A line feed (byte 0x0a).
727 Example: Given appropriate variable definitions, consider the
728 following MRSETS command:
731 MRSETS /MCGROUP NAME=$a LABEL='my mcgroup' VARIABLES=a b c
732 /MDGROUP NAME=$b VARIABLES=g e f d VALUE=55
733 /MDGROUP NAME=$c LABEL='mdgroup #2' VARIABLES=h i j VALUE='Yes'
734 /MDGROUP NAME=$d LABEL='third mdgroup' CATEGORYLABELS=COUNTEDVALUES
735 VARIABLES=k l m VALUE=34
736 /MDGROUP NAME=$e CATEGORYLABELS=COUNTEDVALUES LABELSOURCE=VARLABEL
737 VARIABLES=n o p VALUE='choice'.
740 The above would generate the following multiple response set record of
744 $a=C 10 my mcgroup a b c
746 $c=D3 Yes 10 mdgroup #2 h i j
749 It would also generate the following multiple response set record with
753 $d=E 1 2 34 13 third mdgroup k l m
754 $e=E 11 6 choice 0 n o p
757 @node Variable Display Parameter Record
758 @section Variable Display Parameter Record
760 The variable display parameter record, if present, has the following
770 /* @r{Repeated @code{count} times}. */
772 int32 width; /* @r{Not always present.} */
777 @item int32 rec_type;
778 Record type. Always set to 7.
781 Record subtype. Always set to 11.
784 The size of @code{int32}. Always set to 4.
787 The number of sets of variable display parameters (ordinarily the
788 number of variables in the dictionary), times 2 or 3.
791 The remaining members are repeated @code{count} times, in the same
792 order as the variable records. No element corresponds to variable
793 records that continue long string variables. The meanings of these
794 members are as follows:
798 The measurement type of the variable:
808 SPSS 14 sometimes writes a @code{measure} of 0 for string variables.
809 PSPP interprets this as nominal scale.
812 The width of the display column for the variable in characters.
814 This field is present if @var{count} is 3 times the number of
815 variables in the dictionary. It is omitted if @var{count} is 2 times
816 the number of variables.
818 @item int32 alignment;
819 The alignment of the variable for display purposes:
831 @node Long Variable Names Record
832 @section Long Variable Names Record
834 If present, the long variable names record has the following format:
843 /* @r{Exactly @code{count} bytes of data.} */
844 char var_name_pairs[];
848 @item int32 rec_type;
849 Record type. Always set to 7.
852 Record subtype. Always set to 13.
855 The size of each element in the @code{var_name_pairs} member. Always set to 1.
858 The total number of bytes in @code{var_name_pairs}.
860 @item char var_name_pairs[];
861 A list of @var{key}--@var{value} tuples, where @var{key} is the name
862 of a variable, and @var{value} is its long variable name.
863 The @var{key} field is at most 8 bytes long and must match the
864 name of a variable which appears in the variable record (@pxref{Variable
866 The @var{value} field is at most 64 bytes long.
867 The @var{key} and @var{value} fields are separated by a @samp{=} byte.
868 Each tuple is separated by a byte whose value is 09. There is no
869 trailing separator following the last tuple.
870 The total length is @code{count} bytes.
873 @node Very Long String Record
874 @section Very Long String Record
876 Old versions of SPSS limited string variables to a width of 255 bytes.
877 For backward compatibility with these older versions, the system file
878 format represents a string longer than 255 bytes, called a @dfn{very
879 long string}, as a collection of strings no longer than 255 bytes
880 each. The strings concatenated to make a very long string are called
881 its @dfn{segments}; for consistency, variables other than very long
882 strings are considered to have a single segment.
884 A very long string with a width of @var{w} has @var{n} =
885 (@var{w} + 251) / 252 segments, that is, one segment for every
886 252 bytes of width, rounding up. It would be logical, then, for each
887 of the segments except the last to have a width of 252 and the last
888 segment to have the remainder, but this is not the case. In fact,
889 each segment except the last has a width of 255 bytes. The last
890 segment has width @var{w} - (@var{n} - 1) * 252; some versions
891 of SPSS make it slightly wider, but not wide enough to make the last
892 segment require another 8 bytes of data.
894 Data is packed tightly into segments of a very long string, 255 bytes
895 per segment. Because 255 bytes of segment data are allocated for
896 every 252 bytes of the very long string's width (approximately), some
897 unused space is left over at the end of the allocated segments. Data
898 in unused space is ignored.
900 Example: Consider a very long string of width 20,000. Such a very
901 long string has 20,000 / 252 = 80 (rounding up) segments. The first
902 79 segments have width 255; the last segment has width 20,000 - 79 *
903 252 = 92 or slightly wider (up to 96 bytes, the next multiple of 8).
904 The very long string's data is actually stored in the 19,890 bytes in
905 the first 78 segments, plus the first 110 bytes of the 79th segment
906 (19,890 + 110 = 20,000). The remaining 145 bytes of the 79th segment
907 and all 92 bytes of the 80th segment are unused.
909 The very long string record explains how to stitch together segments
910 to obtain very long string data. For each of the very long string
911 variables in the dictionary, it specifies the name of its first
912 segment's variable and the very long string variable's actual width.
913 The remaining segments immediately follow the named variable in the
914 system file's dictionary.
916 The very long string record, which is present only if the system file
917 contains very long string variables, has the following format:
926 /* @r{Exactly @code{count} bytes of data.} */
927 char string_lengths[];
931 @item int32 rec_type;
932 Record type. Always set to 7.
935 Record subtype. Always set to 14.
938 The size of each element in the @code{string_lengths} member. Always set to 1.
941 The total number of bytes in @code{string_lengths}.
943 @item char string_lengths[];
944 A list of @var{key}--@var{value} tuples, where @var{key} is the name
945 of a variable, and @var{value} is its length.
946 The @var{key} field is at most 8 bytes long and must match the
947 name of a variable which appears in the variable record (@pxref{Variable
949 The @var{value} field is exactly 5 bytes long. It is a zero-padded,
950 ASCII-encoded string that is the length of the variable.
951 The @var{key} and @var{value} fields are separated by a @samp{=} byte.
952 Tuples are delimited by a two-byte sequence @{00, 09@}.
953 After the last tuple, there may be a single byte 00, or @{00, 09@}.
954 The total length is @code{count} bytes.
957 @node Character Encoding Record
958 @section Character Encoding Record
960 This record, if present, indicates the character encoding for string data,
961 long variable names, variable labels, value labels and other strings in the
971 /* @r{Exactly @code{count} bytes of data.} */
976 @item int32 rec_type;
977 Record type. Always set to 7.
980 Record subtype. Always set to 20.
983 The size of each element in the @code{encoding} member. Always set to 1.
986 The total number of bytes in @code{encoding}.
988 @item char encoding[];
989 The name of the character encoding. Normally this will be an official IANA characterset name or alias.
990 See @url{http://www.iana.org/assignments/character-sets}.
993 This record is not present in files generated by older software. See
994 also the @code{character_code} field in the machine integer info
995 record (@pxref{character-code}).
997 When the character encoding record and the machine integer info record
998 are both present, all system files observed in practice indicate the
999 same character encoding, e.g.@: 1252 as @code{character_code} and
1000 @code{windows-1252} as @code{encoding}, 65001 and @code{UTF-8}, etc.
1002 If, for testing purposes, a file is crafted with different
1003 @code{character_code} and @code{encoding}, it seems that
1004 @code{character_code} controls the encoding for all strings in the
1005 system file before the dictionary termination record, including
1006 strings in data (e.g.@: string missing values), and @code{encoding}
1007 controls the encoding for strings following the dictionary termination
1010 @node Long String Value Labels Record
1011 @section Long String Value Labels Record
1013 This record, if present, specifies value labels for long string
1023 /* @r{Repeated up to exactly @code{count} bytes.} */
1028 long_string_label labels[];
1032 @item int32 rec_type;
1033 Record type. Always set to 7.
1035 @item int32 subtype;
1036 Record subtype. Always set to 21.
1042 The number of bytes following the header until the next header.
1044 @item int32 var_name_len;
1045 @itemx char var_name[];
1046 The number of bytes in the name of the variable that has long string
1047 value labels, plus the variable name itself, which consists of exactly
1048 @code{var_name_len} bytes. The variable name is not padded to any
1049 particular boundary, nor is it null-terminated.
1051 @item int32 var_width;
1052 The width of the variable, in bytes, which will be between 9 and
1055 @item int32 n_labels;
1056 @itemx long_string_label labels[];
1057 The long string labels themselves. The @code{labels} array contains
1058 exactly @code{n_labels} elements, each of which has the following
1069 @item int32 value_len;
1070 @itemx char value[];
1071 The string value being labeled. @code{value_len} is the number of
1072 bytes in @code{value}; it is equal to @code{var_width}. The
1073 @code{value} array is not padded or null-terminated.
1075 @item int32 label_len;
1076 @itemx char label[];
1077 The label for the string value. @code{label_len}, which must be
1078 between 0 and 120, is the number of bytes in @code{label}. The
1079 @code{label} array is not padded or null-terminated.
1083 @node Data File and Variable Attributes Records
1084 @section Data File and Variable Attributes Records
1086 The data file and variable attributes records represent custom
1087 attributes for the system file or for individual variables in the
1088 system file, as defined on the DATAFILE ATTRIBUTE (@pxref{DATAFILE
1089 ATTRIBUTE,,,pspp, PSPP Users Guide}) and VARIABLE ATTRIBUTE commands
1090 (@pxref{VARIABLE ATTRIBUTE,,,pspp, PSPP Users Guide}), respectively.
1099 /* @r{Exactly @code{count} bytes of data.} */
1104 @item int32 rec_type;
1105 Record type. Always set to 7.
1107 @item int32 subtype;
1108 Record subtype. Always set to 17 for a data file attribute record or
1109 to 18 for a variable attributes record.
1112 The size of each element in the @code{attributes} member. Always set to 1.
1115 The total number of bytes in @code{attributes}.
1117 @item char attributes[];
1118 The attributes, in a text-based format.
1120 In record type 17, this field contains a single attribute set. An
1121 attribute set is a sequence of one or more attributes concatenated
1122 together. Each attribute consists of a name, which has the same
1123 syntax as a variable name, followed by, inside parentheses, a sequence
1124 of one or more values. Each value consists of a string enclosed in
1125 single quotes (@code{'}) followed by a line feed (byte 0x0a). A value
1126 may contain single quote characters, which are not themselves escaped
1127 or quoted or required to be present in pairs. There is no apparent
1128 way to embed a line feed in a value. There is no distinction between
1129 an attribute with a single value and an attribute array with one
1132 In record type 18, this field contains a sequence of one or more
1133 variable attribute sets. If more than one variable attribute set is
1134 present, each one after the first is delimited from the previous by
1135 @code{/}. Each variable attribute set consists of a (potentially
1136 long) variable name,
1137 followed by @code{:}, followed by an attribute set with the same
1138 syntax as on record type 17.
1140 The total length is @code{count} bytes.
1145 A system file produced with the following VARIABLE ATTRIBUTE commands
1149 VARIABLE ATTRIBUTE VARIABLES=dummy ATTRIBUTE=fred[1]('23') fred[2]('34').
1150 VARIABLE ATTRIBUTE VARIABLES=dummy ATTRIBUTE=bert('123').
1154 will contain a variable attribute record with the following contents:
1157 00000000 07 00 00 00 12 00 00 00 01 00 00 00 22 00 00 00 |............"...|
1158 00000010 64 75 6d 6d 79 3a 66 72 65 64 28 27 32 33 27 0a |dummy:fred('23'.|
1159 00000020 27 33 34 27 0a 29 62 65 72 74 28 27 31 32 33 27 |'34'.)bert('123'|
1160 00000030 0a 29 |.) |
1163 @node Extended Number of Cases Record
1164 @section Extended Number of Cases Record
1166 The file header record expresses the number of cases in the system
1167 file as an int32 (@pxref{File Header Record}). This record allows the
1168 number of cases in the system file to be expressed as a 64-bit number.
1180 @item int32 rec_type;
1181 Record type. Always set to 7.
1183 @item int32 subtype;
1184 Record subtype. Always set to 16.
1187 Size of each element. Always set to 8.
1190 Number of pieces of data in the data part. Alway set to 2.
1192 @item int64 unknown;
1193 Meaning unknown. Always set to 1.
1195 @item int64 ncases64;
1196 Number of cases in the file as a 64-bit integer. Presumably this
1197 could be -1 to indicate that the number of cases is unknown, for the
1198 same reason as @code{ncases} in the file header record, but this has
1199 not been observed in the wild.
1202 @node Miscellaneous Informational Records
1203 @section Miscellaneous Informational Records
1205 Some specific types of miscellaneous informational records are
1206 documented here, but others are known to exist. PSPP ignores unknown
1207 miscellaneous informational records when reading system files.
1216 /* @r{Exactly @code{size * count} bytes of data.} */
1221 @item int32 rec_type;
1222 Record type. Always set to 7.
1224 @item int32 subtype;
1225 Record subtype. May take any value. According to Aapi
1226 H@"am@"al@"ainen, value 5 indicates a set of grouped variables and 6
1227 indicates date info (probably related to USE).
1230 Size of each piece of data in the data part. Should have the value 1,
1231 4, or 8, for @code{char}, @code{int32}, and @code{flt64} format data,
1235 Number of pieces of data in the data part.
1238 Arbitrary data. There must be @code{size} times @code{count} bytes of
1242 @node Dictionary Termination Record
1243 @section Dictionary Termination Record
1245 The dictionary termination record separates all other records from the
1254 @item int32 rec_type;
1255 Record type. Always set to 999.
1258 Ignored padding. Should be set to 0.
1262 @section Data Record
1264 Data records must follow all other records in the system file. There must
1265 be at least one data record in every system file.
1267 The format of data records varies depending on whether the data is
1268 compressed. Regardless, the data is arranged in a series of 8-byte
1271 When data is not compressed,
1272 each element corresponds to
1273 the variable declared in the respective variable record (@pxref{Variable
1274 Record}). Numeric values are given in @code{flt64} format; string
1275 values are literal characters string, padded on the right when
1276 necessary to fill out 8-byte units.
1278 Compressed data is arranged in the following manner: the first 8 bytes
1279 in the data section is divided into a series of 1-byte command
1280 codes. These codes have meanings as described below:
1284 Ignored. If the program writing the system file accumulates compressed
1285 data in blocks of fixed length, 0 bytes can be used to pad out extra
1286 bytes remaining at the end of a fixed-size block.
1290 value @var{code} - @var{bias}, where
1291 @var{code} is the value of the compression code and @var{bias} is the
1292 variable @code{bias} from the file header. For example,
1293 code 105 with bias 100.0 (the normal value) indicates a numeric variable
1295 One file has been seen written by SPSS 14 that contained such a code
1296 in a @emph{string} field with the value 0 (after the bias is
1297 subtracted) as a way of encoding null bytes.
1300 End of file. This code may or may not appear at the end of the data
1301 stream. PSPP always outputs this code but its use is not required.
1304 A numeric or string value that is not
1305 compressible. The value is stored in the 8 bytes following the
1306 current block of command bytes. If this value appears twice in a block
1307 of command bytes, then it indicates the second group of 8 bytes following the
1308 command bytes, and so on.
1311 An 8-byte string value that is all spaces.
1314 The system-missing value.
1317 When the end of the an 8-byte group of command bytes is reached, any
1318 blocks of non-compressible values indicated by code 253 are skipped,
1319 and the next element of command bytes is read and interpreted, until
1320 the end of the file or a code with value 252 is reached.
1321 @setfilename ignored