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 Extension (type 7) records, in ascending numerical order of their
67 Dictionary termination record.
73 Each type of record is described separately below.
76 * File Header Record::
78 * Value Labels Records::
80 * Machine Integer Info Record::
81 * Machine Floating-Point Info Record::
82 * Multiple Response Sets Records::
83 * Variable Display Parameter Record::
84 * Long Variable Names Record::
85 * Very Long String Record::
86 * Character Encoding Record::
87 * Long String Value Labels Record::
88 * Data File and Variable Attributes Records::
89 * Extended Number of Cases Record::
90 * Miscellaneous Informational Records::
91 * Dictionary Termination Record::
95 @node File Header Record
96 @section File Header Record
98 The file header is always the first record in the file. It has the
105 int32 nominal_case_size;
110 char creation_date[9];
111 char creation_time[8];
117 @item char rec_type[4];
118 Record type code, set to @samp{$FL2}.
120 @item char prod_name[60];
121 Product identification string. This always begins with the characters
122 @samp{@@(#) SPSS DATA FILE}. PSPP uses the remaining characters to
123 give its version and the operating system name; for example, @samp{GNU
124 pspp 0.1.4 - sparc-sun-solaris2.5.2}. The string is truncated if it
125 would be longer than 60 characters; otherwise it is padded on the right
129 @item int32 layout_code;
130 Normally set to 2, although a few system files have been spotted in
131 the wild with a value of 3 here. PSPP use this value to determine the
132 file's integer endianness (@pxref{System File Format}).
134 @item int32 nominal_case_size;
135 Number of data elements per case. This is the number of variables,
136 except that long string variables add extra data elements (one for every
137 8 characters after the first 8). However, string variables do not
138 contribute to this value beyond the first 255 bytes. Further, system
139 files written by some systems set this value to -1. In general, it is
140 unsafe for systems reading system files to rely upon this value.
142 @item int32 compressed;
143 Set to 1 if the data in the file is compressed, 0 otherwise.
145 @item int32 weight_index;
146 If one of the variables in the data set is used as a weighting
147 variable, set to the dictionary index of that variable, plus 1
148 (@pxref{Dictionary Index}). Otherwise, set to 0.
151 Set to the number of cases in the file if it is known, or -1 otherwise.
153 In the general case it is not possible to determine the number of cases
154 that will be output to a system file at the time that the header is
155 written. The way that this is dealt with is by writing the entire
156 system file, including the header, then seeking back to the beginning of
157 the file and writing just the @code{ncases} field. For files in which
158 this is not valid, the seek operation fails. In this case,
159 @code{ncases} remains -1.
163 Compression bias, ordinarily set to 100. Only integers between
164 @code{1 - bias} and @code{251 - bias} can be compressed.
166 By assuming that its value is 100, PSPP uses @code{bias} to determine
167 the file's floating-point format and endianness (@pxref{System File
168 Format}). If the compression bias is not 100, PSPP cannot auto-detect
169 the floating-point format and assumes that it is IEEE 754 format with
170 the same endianness as the system file's integers, which is correct
171 for all known system files.
173 @item char creation_date[9];
174 Date of creation of the system file, in @samp{dd mmm yy}
175 format, with the month as standard English abbreviations, using an
176 initial capital letter and following with lowercase. If the date is not
177 available then this field is arbitrarily set to @samp{01 Jan 70}.
179 @item char creation_time[8];
180 Time of creation of the system file, in @samp{hh:mm:ss}
181 format and using 24-hour time. If the time is not available then this
182 field is arbitrarily set to @samp{00:00:00}.
184 @item char file_label[64];
185 File label declared by the user, if any (@pxref{FILE LABEL,,,pspp,
186 PSPP Users Guide}). Padded on the right with spaces.
188 @item char padding[3];
189 Ignored padding bytes to make the structure a multiple of 32 bits in
190 length. Set to zeros.
193 @node Variable Record
194 @section Variable Record
196 There must be one variable record for each numeric variable and each
197 string variable with width 8 bytes or less. String variables wider
198 than 8 bytes have one variable record for each 8 bytes, rounding up.
199 The first variable record for a long string specifies the variable's
200 correct dictionary information. Subsequent variable records for a
201 long string are filled with dummy information: a type of -1, no
202 variable label or missing values, print and write formats that are
203 ignored, and an empty string as name. A few system files have been
204 encountered that include a variable label on dummy variable records,
205 so readers should take care to parse dummy variable records in the
206 same way as other variable records.
208 @anchor{Dictionary Index}
209 The @dfn{dictionary index} of a variable is its offset in the set of
210 variable records, including dummy variable records for long string
211 variables. The first variable record has a dictionary index of 0, the
212 second has a dictionary index of 1, and so on.
214 The system file format does not directly support string variables
215 wider than 255 bytes. Such very long string variables are represented
216 by a number of narrower string variables. @xref{Very Long String
217 Record}, for details.
223 int32 n_missing_values;
228 /* @r{Present only if @code{has_var_label} is 1.} */
232 /* @r{Present only if @code{n_missing_values} is nonzero}. */
233 flt64 missing_values[];
237 @item int32 rec_type;
238 Record type code. Always set to 2.
241 Variable type code. Set to 0 for a numeric variable. For a short
242 string variable or the first part of a long string variable, this is set
243 to the width of the string. For the second and subsequent parts of a
244 long string variable, set to -1, and the remaining fields in the
245 structure are ignored.
247 @item int32 has_var_label;
248 If this variable has a variable label, set to 1; otherwise, set to 0.
250 @item int32 n_missing_values;
251 If the variable has no missing values, set to 0. If the variable has
252 one, two, or three discrete missing values, set to 1, 2, or 3,
253 respectively. If the variable has a range for missing variables, set to
254 -2; if the variable has a range for missing variables plus a single
255 discrete value, set to -3.
258 Print format for this variable. See below.
261 Write format for this variable. See below.
264 Variable name. The variable name must begin with a capital letter or
265 the at-sign (@samp{@@}). Subsequent characters may also be digits, octothorpes
266 (@samp{#}), dollar signs (@samp{$}), underscores (@samp{_}), or full
267 stops (@samp{.}). The variable name is padded on the right with spaces.
269 @item int32 label_len;
270 This field is present only if @code{has_var_label} is set to 1. It is
271 set to the length, in characters, of the variable label. The
272 documented maximum length varies from 120 to 255 based on SPSS
273 version, but some files have been seen with longer labels. PSPP
274 accepts longer labels and truncates them to 255 bytes on input.
277 This field is present only if @code{has_var_label} is set to 1. It has
278 length @code{label_len}, rounded up to the nearest multiple of 32 bits.
279 The first @code{label_len} characters are the variable's variable label.
281 @item flt64 missing_values[];
282 This field is present only if @code{n_missing_values} is nonzero. It
283 has the same number of 8-byte elements as the absolute value of
284 @code{n_missing_values}. Each element is interpreted as a number for
285 numeric variables (with HIGHEST and LOWEST indicated as described in
286 the chapter introduction). For string variables of width less than 8
287 bytes, elements are right-padded with spaces; for string variables
288 wider than 8 bytes, only the first 8 bytes of each missing value are
289 specified, with the remainder implicitly all spaces.
291 For discrete missing values, each element represents one missing
292 value. When a range is present, the first element denotes the minimum
293 value in the range, and the second element denotes the maximum value
294 in the range. When a range plus a value are present, the third
295 element denotes the additional discrete missing value.
298 The @code{print} and @code{write} members of sysfile_variable are output
299 formats coded into @code{int32} types. The least-significant byte
300 of the @code{int32} represents the number of decimal places, and the
301 next two bytes in order of increasing significance represent field width
302 and format type, respectively. The most-significant byte is not
303 used and should be set to zero.
305 Format types are defined as follows:
308 @multitable {Value} {@code{DATETIME}}
394 @node Value Labels Records
395 @section Value Labels Records
397 The value label records documented in this section are used for
398 numeric and short string variables only. Long string variables may
399 have value labels, but their value labels are recorded using a
400 different record type (@pxref{Long String Value Labels Record}).
402 The value label record has the following format:
408 /* @r{Repeated @code{label_cnt} times}. */
415 @item int32 rec_type;
416 Record type. Always set to 3.
418 @item int32 label_count;
419 Number of value labels present in this record.
422 The remaining fields are repeated @code{count} times. Each
423 repetition specifies one value label.
427 A numeric value or a short string value padded as necessary to 8 bytes
428 in length. Its type and width cannot be determined until the
429 following value label variables record (see below) is read.
431 @item char label_len;
432 The label's length, in bytes. The documented maximum length varies
433 from 60 to 120 based on SPSS version. PSPP supports value labels up
437 @code{label_len} bytes of the actual label, followed by up to 7 bytes
438 of padding to bring @code{label} and @code{label_len} together to a
439 multiple of 8 bytes in length.
442 The value label record is always immediately followed by a value label
443 variables record with the following format:
452 @item int32 rec_type;
453 Record type. Always set to 4.
455 @item int32 var_count;
456 Number of variables that the associated value labels from the value
457 label record are to be applied.
460 A list of dictionary indexes of variables to which to apply the value
461 labels (@pxref{Dictionary Index}). There are @code{var_count}
464 String variables wider than 8 bytes may not be specified in this list.
467 @node Document Record
468 @section Document Record
470 The document record, if present, has the following format:
479 @item int32 rec_type;
480 Record type. Always set to 6.
483 Number of lines of documents present.
485 @item char lines[][80];
486 Document lines. The number of elements is defined by @code{n_lines}.
487 Lines shorter than 80 characters are padded on the right with spaces.
490 @node Machine Integer Info Record
491 @section Machine Integer Info Record
493 The integer info record, if present, has the following format:
505 int32 version_revision;
507 int32 floating_point_rep;
508 int32 compression_code;
510 int32 character_code;
514 @item int32 rec_type;
515 Record type. Always set to 7.
518 Record subtype. Always set to 3.
521 Size of each piece of data in the data part, in bytes. Always set to 4.
524 Number of pieces of data in the data part. Always set to 8.
526 @item int32 version_major;
527 PSPP major version number. In version @var{x}.@var{y}.@var{z}, this
530 @item int32 version_minor;
531 PSPP minor version number. In version @var{x}.@var{y}.@var{z}, this
534 @item int32 version_revision;
535 PSPP version revision number. In version @var{x}.@var{y}.@var{z},
538 @item int32 machine_code;
539 Machine code. PSPP always set this field to value to -1, but other
542 @item int32 floating_point_rep;
543 Floating point representation code. For IEEE 754 systems this is 1.
544 IBM 370 sets this to 2, and DEC VAX E to 3.
546 @item int32 compression_code;
547 Compression code. Always set to 1.
549 @item int32 endianness;
550 Machine endianness. 1 indicates big-endian, 2 indicates little-endian.
552 @item int32 character_code;
553 @anchor{character-code} Character code. The following values have
554 been actually observed in system files:
561 The @code{windows-1250} code page for Central European and Eastern
565 The @code{windows-1252} code page for Western European languages.
574 The following additional values are known to be defined:
587 Other Windows code page numbers are known to be generally valid.
589 Old versions of SPSS always wrote value 2 in this field, regardless of
590 the encoding in use. Newer versions also write the character encoding
591 as a string (see @ref{Character Encoding Record}).
594 @node Machine Floating-Point Info Record
595 @section Machine Floating-Point Info Record
597 The floating-point info record, if present, has the following format:
613 @item int32 rec_type;
614 Record type. Always set to 7.
617 Record subtype. Always set to 4.
620 Size of each piece of data in the data part, in bytes. Always set to 8.
623 Number of pieces of data in the data part. Always set to 3.
626 The system missing value.
629 The value used for HIGHEST in missing values.
632 The value used for LOWEST in missing values.
635 @node Multiple Response Sets Records
636 @section Multiple Response Sets Records
638 The system file format has two different types of records that
639 represent multiple response sets (@pxref{MRSETS,,,pspp, PSPP Users
640 Guide}). The first type of record describes multiple response sets
641 that can be understood by SPSS before version 14. The second type of
642 record, with a closely related format, is used for multiple dichotomy
643 sets that use the CATEGORYLABELS=COUNTEDVALUES feature added in
653 /* @r{Exactly @code{count} bytes of data.} */
658 @item int32 rec_type;
659 Record type. Always set to 7.
662 Record subtype. Set to 7 for records that describe multiple response
663 sets understood by SPSS before version 14, or to 19 for records that
664 describe dichotomy sets that use the CATEGORYLABELS=COUNTEDVALUES
665 feature added in version 14.
668 The size of each element in the @code{mrsets} member. Always set to 1.
671 The total number of bytes in @code{mrsets}.
674 A series of multiple response sets, each of which consists of the
679 The set's name (an identifier that begins with @samp{$}).
682 An equals sign (@samp{=}).
685 @samp{C} for a multiple category set, @samp{D} for a multiple
686 dichotomy set with CATEGORYLABELS=VARLABELS, or @samp{E} for a
687 multiple dichotomy set with CATEGORYLABELS=COUNTEDVALUES.
690 For a multiple dichotomy set with CATEGORYLABELS=COUNTEDVALUES, a
691 space, followed by a number expressed as decimal digits, followed by a
692 space. If LABELSOURCE=VARLABEL was specified on MRSETS, then the
693 number is 11; otherwise it is 1.@footnote{This part of the format may
694 not be fully understood, because only a single example of each
695 possibility has been examined.}
698 For either kind of multiple dichotomy set, the counted value, as a
699 positive integer count specified as decimal digits, followed by a
700 space, followed by as many string bytes as specified in the count. If
701 the set contains numeric variables, the string consists of the counted
702 integer value expressed as decimal digits. If the set contains string
703 variables, the string contains the counted string value. Either way,
704 the string may be padded on the right with spaces (older versions of
705 SPSS seem to always pad to a width of 8 bytes; newer versions don't).
711 The multiple response set's label, using the same format as for the
712 counted value for multiple dichotomy sets. A string of length 0 means
713 that the set does not have a label. A string of length 0 is also
714 written if LABELSOURCE=VARLABEL was specified.
720 The names of the variables in the set, each separated from the
721 previous by a single space.
724 A line feed (byte 0x0a).
728 Example: Given appropriate variable definitions, consider the
729 following MRSETS command:
732 MRSETS /MCGROUP NAME=$a LABEL='my mcgroup' VARIABLES=a b c
733 /MDGROUP NAME=$b VARIABLES=g e f d VALUE=55
734 /MDGROUP NAME=$c LABEL='mdgroup #2' VARIABLES=h i j VALUE='Yes'
735 /MDGROUP NAME=$d LABEL='third mdgroup' CATEGORYLABELS=COUNTEDVALUES
736 VARIABLES=k l m VALUE=34
737 /MDGROUP NAME=$e CATEGORYLABELS=COUNTEDVALUES LABELSOURCE=VARLABEL
738 VARIABLES=n o p VALUE='choice'.
741 The above would generate the following multiple response set record of
745 $a=C 10 my mcgroup a b c
747 $c=D3 Yes 10 mdgroup #2 h i j
750 It would also generate the following multiple response set record with
754 $d=E 1 2 34 13 third mdgroup k l m
755 $e=E 11 6 choice 0 n o p
758 @node Variable Display Parameter Record
759 @section Variable Display Parameter Record
761 The variable display parameter record, if present, has the following
771 /* @r{Repeated @code{count} times}. */
773 int32 width; /* @r{Not always present.} */
778 @item int32 rec_type;
779 Record type. Always set to 7.
782 Record subtype. Always set to 11.
785 The size of @code{int32}. Always set to 4.
788 The number of sets of variable display parameters (ordinarily the
789 number of variables in the dictionary), times 2 or 3.
792 The remaining members are repeated @code{count} times, in the same
793 order as the variable records. No element corresponds to variable
794 records that continue long string variables. The meanings of these
795 members are as follows:
799 The measurement type of the variable:
809 SPSS 14 sometimes writes a @code{measure} of 0 for string variables.
810 PSPP interprets this as nominal scale.
813 The width of the display column for the variable in characters.
815 This field is present if @var{count} is 3 times the number of
816 variables in the dictionary. It is omitted if @var{count} is 2 times
817 the number of variables.
819 @item int32 alignment;
820 The alignment of the variable for display purposes:
832 @node Long Variable Names Record
833 @section Long Variable Names Record
835 If present, the long variable names record has the following format:
844 /* @r{Exactly @code{count} bytes of data.} */
845 char var_name_pairs[];
849 @item int32 rec_type;
850 Record type. Always set to 7.
853 Record subtype. Always set to 13.
856 The size of each element in the @code{var_name_pairs} member. Always set to 1.
859 The total number of bytes in @code{var_name_pairs}.
861 @item char var_name_pairs[];
862 A list of @var{key}--@var{value} tuples, where @var{key} is the name
863 of a variable, and @var{value} is its long variable name.
864 The @var{key} field is at most 8 bytes long and must match the
865 name of a variable which appears in the variable record (@pxref{Variable
867 The @var{value} field is at most 64 bytes long.
868 The @var{key} and @var{value} fields are separated by a @samp{=} byte.
869 Each tuple is separated by a byte whose value is 09. There is no
870 trailing separator following the last tuple.
871 The total length is @code{count} bytes.
874 @node Very Long String Record
875 @section Very Long String Record
877 Old versions of SPSS limited string variables to a width of 255 bytes.
878 For backward compatibility with these older versions, the system file
879 format represents a string longer than 255 bytes, called a @dfn{very
880 long string}, as a collection of strings no longer than 255 bytes
881 each. The strings concatenated to make a very long string are called
882 its @dfn{segments}; for consistency, variables other than very long
883 strings are considered to have a single segment.
885 A very long string with a width of @var{w} has @var{n} =
886 (@var{w} + 251) / 252 segments, that is, one segment for every
887 252 bytes of width, rounding up. It would be logical, then, for each
888 of the segments except the last to have a width of 252 and the last
889 segment to have the remainder, but this is not the case. In fact,
890 each segment except the last has a width of 255 bytes. The last
891 segment has width @var{w} - (@var{n} - 1) * 252; some versions
892 of SPSS make it slightly wider, but not wide enough to make the last
893 segment require another 8 bytes of data.
895 Data is packed tightly into segments of a very long string, 255 bytes
896 per segment. Because 255 bytes of segment data are allocated for
897 every 252 bytes of the very long string's width (approximately), some
898 unused space is left over at the end of the allocated segments. Data
899 in unused space is ignored.
901 Example: Consider a very long string of width 20,000. Such a very
902 long string has 20,000 / 252 = 80 (rounding up) segments. The first
903 79 segments have width 255; the last segment has width 20,000 - 79 *
904 252 = 92 or slightly wider (up to 96 bytes, the next multiple of 8).
905 The very long string's data is actually stored in the 19,890 bytes in
906 the first 78 segments, plus the first 110 bytes of the 79th segment
907 (19,890 + 110 = 20,000). The remaining 145 bytes of the 79th segment
908 and all 92 bytes of the 80th segment are unused.
910 The very long string record explains how to stitch together segments
911 to obtain very long string data. For each of the very long string
912 variables in the dictionary, it specifies the name of its first
913 segment's variable and the very long string variable's actual width.
914 The remaining segments immediately follow the named variable in the
915 system file's dictionary.
917 The very long string record, which is present only if the system file
918 contains very long string variables, has the following format:
927 /* @r{Exactly @code{count} bytes of data.} */
928 char string_lengths[];
932 @item int32 rec_type;
933 Record type. Always set to 7.
936 Record subtype. Always set to 14.
939 The size of each element in the @code{string_lengths} member. Always set to 1.
942 The total number of bytes in @code{string_lengths}.
944 @item char string_lengths[];
945 A list of @var{key}--@var{value} tuples, where @var{key} is the name
946 of a variable, and @var{value} is its length.
947 The @var{key} field is at most 8 bytes long and must match the
948 name of a variable which appears in the variable record (@pxref{Variable
950 The @var{value} field is exactly 5 bytes long. It is a zero-padded,
951 ASCII-encoded string that is the length of the variable.
952 The @var{key} and @var{value} fields are separated by a @samp{=} byte.
953 Tuples are delimited by a two-byte sequence @{00, 09@}.
954 After the last tuple, there may be a single byte 00, or @{00, 09@}.
955 The total length is @code{count} bytes.
958 @node Character Encoding Record
959 @section Character Encoding Record
961 This record, if present, indicates the character encoding for string data,
962 long variable names, variable labels, value labels and other strings in the
972 /* @r{Exactly @code{count} bytes of data.} */
977 @item int32 rec_type;
978 Record type. Always set to 7.
981 Record subtype. Always set to 20.
984 The size of each element in the @code{encoding} member. Always set to 1.
987 The total number of bytes in @code{encoding}.
989 @item char encoding[];
990 The name of the character encoding. Normally this will be an official IANA characterset name or alias.
991 See @url{http://www.iana.org/assignments/character-sets}.
994 This record is not present in files generated by older software. See
995 also the @code{character_code} field in the machine integer info
996 record (@pxref{character-code}).
998 When the character encoding record and the machine integer info record
999 are both present, all system files observed in practice indicate the
1000 same character encoding, e.g.@: 1252 as @code{character_code} and
1001 @code{windows-1252} as @code{encoding}, 65001 and @code{UTF-8}, etc.
1003 If, for testing purposes, a file is crafted with different
1004 @code{character_code} and @code{encoding}, it seems that
1005 @code{character_code} controls the encoding for all strings in the
1006 system file before the dictionary termination record, including
1007 strings in data (e.g.@: string missing values), and @code{encoding}
1008 controls the encoding for strings following the dictionary termination
1011 @node Long String Value Labels Record
1012 @section Long String Value Labels Record
1014 This record, if present, specifies value labels for long string
1024 /* @r{Repeated up to exactly @code{count} bytes.} */
1029 long_string_label labels[];
1033 @item int32 rec_type;
1034 Record type. Always set to 7.
1036 @item int32 subtype;
1037 Record subtype. Always set to 21.
1043 The number of bytes following the header until the next header.
1045 @item int32 var_name_len;
1046 @itemx char var_name[];
1047 The number of bytes in the name of the variable that has long string
1048 value labels, plus the variable name itself, which consists of exactly
1049 @code{var_name_len} bytes. The variable name is not padded to any
1050 particular boundary, nor is it null-terminated.
1052 @item int32 var_width;
1053 The width of the variable, in bytes, which will be between 9 and
1056 @item int32 n_labels;
1057 @itemx long_string_label labels[];
1058 The long string labels themselves. The @code{labels} array contains
1059 exactly @code{n_labels} elements, each of which has the following
1070 @item int32 value_len;
1071 @itemx char value[];
1072 The string value being labeled. @code{value_len} is the number of
1073 bytes in @code{value}; it is equal to @code{var_width}. The
1074 @code{value} array is not padded or null-terminated.
1076 @item int32 label_len;
1077 @itemx char label[];
1078 The label for the string value. @code{label_len}, which must be
1079 between 0 and 120, is the number of bytes in @code{label}. The
1080 @code{label} array is not padded or null-terminated.
1084 @node Data File and Variable Attributes Records
1085 @section Data File and Variable Attributes Records
1087 The data file and variable attributes records represent custom
1088 attributes for the system file or for individual variables in the
1089 system file, as defined on the DATAFILE ATTRIBUTE (@pxref{DATAFILE
1090 ATTRIBUTE,,,pspp, PSPP Users Guide}) and VARIABLE ATTRIBUTE commands
1091 (@pxref{VARIABLE ATTRIBUTE,,,pspp, PSPP Users Guide}), respectively.
1100 /* @r{Exactly @code{count} bytes of data.} */
1105 @item int32 rec_type;
1106 Record type. Always set to 7.
1108 @item int32 subtype;
1109 Record subtype. Always set to 17 for a data file attribute record or
1110 to 18 for a variable attributes record.
1113 The size of each element in the @code{attributes} member. Always set to 1.
1116 The total number of bytes in @code{attributes}.
1118 @item char attributes[];
1119 The attributes, in a text-based format.
1121 In record type 17, this field contains a single attribute set. An
1122 attribute set is a sequence of one or more attributes concatenated
1123 together. Each attribute consists of a name, which has the same
1124 syntax as a variable name, followed by, inside parentheses, a sequence
1125 of one or more values. Each value consists of a string enclosed in
1126 single quotes (@code{'}) followed by a line feed (byte 0x0a). A value
1127 may contain single quote characters, which are not themselves escaped
1128 or quoted or required to be present in pairs. There is no apparent
1129 way to embed a line feed in a value. There is no distinction between
1130 an attribute with a single value and an attribute array with one
1133 In record type 18, this field contains a sequence of one or more
1134 variable attribute sets. If more than one variable attribute set is
1135 present, each one after the first is delimited from the previous by
1136 @code{/}. Each variable attribute set consists of a (potentially
1137 long) variable name,
1138 followed by @code{:}, followed by an attribute set with the same
1139 syntax as on record type 17.
1141 The total length is @code{count} bytes.
1146 A system file produced with the following VARIABLE ATTRIBUTE commands
1150 VARIABLE ATTRIBUTE VARIABLES=dummy ATTRIBUTE=fred[1]('23') fred[2]('34').
1151 VARIABLE ATTRIBUTE VARIABLES=dummy ATTRIBUTE=bert('123').
1155 will contain a variable attribute record with the following contents:
1158 00000000 07 00 00 00 12 00 00 00 01 00 00 00 22 00 00 00 |............"...|
1159 00000010 64 75 6d 6d 79 3a 66 72 65 64 28 27 32 33 27 0a |dummy:fred('23'.|
1160 00000020 27 33 34 27 0a 29 62 65 72 74 28 27 31 32 33 27 |'34'.)bert('123'|
1161 00000030 0a 29 |.) |
1164 @node Extended Number of Cases Record
1165 @section Extended Number of Cases Record
1167 The file header record expresses the number of cases in the system
1168 file as an int32 (@pxref{File Header Record}). This record allows the
1169 number of cases in the system file to be expressed as a 64-bit number.
1181 @item int32 rec_type;
1182 Record type. Always set to 7.
1184 @item int32 subtype;
1185 Record subtype. Always set to 16.
1188 Size of each element. Always set to 8.
1191 Number of pieces of data in the data part. Alway set to 2.
1193 @item int64 unknown;
1194 Meaning unknown. Always set to 1.
1196 @item int64 ncases64;
1197 Number of cases in the file as a 64-bit integer. Presumably this
1198 could be -1 to indicate that the number of cases is unknown, for the
1199 same reason as @code{ncases} in the file header record, but this has
1200 not been observed in the wild.
1203 @node Miscellaneous Informational Records
1204 @section Miscellaneous Informational Records
1206 Some specific types of miscellaneous informational records are
1207 documented here, but others are known to exist. PSPP ignores unknown
1208 miscellaneous informational records when reading system files.
1217 /* @r{Exactly @code{size * count} bytes of data.} */
1222 @item int32 rec_type;
1223 Record type. Always set to 7.
1225 @item int32 subtype;
1226 Record subtype. May take any value. According to Aapi
1227 H@"am@"al@"ainen, value 5 indicates a set of grouped variables and 6
1228 indicates date info (probably related to USE).
1231 Size of each piece of data in the data part. Should have the value 1,
1232 4, or 8, for @code{char}, @code{int32}, and @code{flt64} format data,
1236 Number of pieces of data in the data part.
1239 Arbitrary data. There must be @code{size} times @code{count} bytes of
1243 @node Dictionary Termination Record
1244 @section Dictionary Termination Record
1246 The dictionary termination record separates all other records from the
1255 @item int32 rec_type;
1256 Record type. Always set to 999.
1259 Ignored padding. Should be set to 0.
1263 @section Data Record
1265 Data records must follow all other records in the system file. There must
1266 be at least one data record in every system file.
1268 The format of data records varies depending on whether the data is
1269 compressed. Regardless, the data is arranged in a series of 8-byte
1272 When data is not compressed,
1273 each element corresponds to
1274 the variable declared in the respective variable record (@pxref{Variable
1275 Record}). Numeric values are given in @code{flt64} format; string
1276 values are literal characters string, padded on the right when
1277 necessary to fill out 8-byte units.
1279 Compressed data is arranged in the following manner: the first 8 bytes
1280 in the data section is divided into a series of 1-byte command
1281 codes. These codes have meanings as described below:
1285 Ignored. If the program writing the system file accumulates compressed
1286 data in blocks of fixed length, 0 bytes can be used to pad out extra
1287 bytes remaining at the end of a fixed-size block.
1291 value @var{code} - @var{bias}, where
1292 @var{code} is the value of the compression code and @var{bias} is the
1293 variable @code{bias} from the file header. For example,
1294 code 105 with bias 100.0 (the normal value) indicates a numeric variable
1296 One file has been seen written by SPSS 14 that contained such a code
1297 in a @emph{string} field with the value 0 (after the bias is
1298 subtracted) as a way of encoding null bytes.
1301 End of file. This code may or may not appear at the end of the data
1302 stream. PSPP always outputs this code but its use is not required.
1305 A numeric or string value that is not
1306 compressible. The value is stored in the 8 bytes following the
1307 current block of command bytes. If this value appears twice in a block
1308 of command bytes, then it indicates the second group of 8 bytes following the
1309 command bytes, and so on.
1312 An 8-byte string value that is all spaces.
1315 The system-missing value.
1318 When the end of the an 8-byte group of command bytes is reached, any
1319 blocks of non-compressible values indicated by code 253 are skipped,
1320 and the next element of command bytes is read and interpreted, until
1321 the end of the file or a code with value 252 is reached.
1322 @setfilename ignored