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 System files use a few floating point values for special purposes:
40 The system-missing value is represented by the largest possible
41 negative number in the floating point format (@code{-DBL_MAX}).
44 HIGHEST is used as the high end of a missing value range with an
45 unbounded maximum. It is represented by the largest possible positive
46 number (@code{DBL_MAX}).
49 LOWEST is used as the low end of a missing value range with an
50 unbounded minimum. It was originally represented by the
51 second-largest negative number (in IEEE 754 format,
52 @code{0xffeffffffffffffe}). System files written by SPSS 21 and later
53 instead use the largest negative number (@code{-DBL_MAX}), the same
54 value as SYSMIS. This does not lead to ambiguity because LOWEST
55 appears in system files only in missing value ranges, which never
59 System files may use most character encodings based on an 8-bit unit.
60 UTF-16 and UTF-32, based on wider units, appear to be unacceptable.
61 @code{rec_type} in the file header record is sufficient to distinguish
62 between ASCII and EBCDIC based encodings. The best way to determine
63 the specific encoding in use is to consult the character encoding
64 record (@pxref{Character Encoding Record}), if present, and failing
65 that the @code{character_code} in the machine integer info record
66 (@pxref{Machine Integer Info Record}). The same encoding should be
67 used for the dictionary and the data in the file, although it is
68 possible to artificially synthesize files that use different encodings
69 (@pxref{Character Encoding Record}).
71 System files are divided into records, each of which begins with a
72 4-byte record type, usually regarded as an @code{int32}.
74 The records must appear in the following order:
84 All pairs of value labels records and value label variables records,
88 Document record, if present.
91 Extension (type 7) records, in ascending numerical order of their
95 Dictionary termination record.
101 Each type of record is described separately below.
104 * File Header Record::
106 * Value Labels Records::
108 * Machine Integer Info Record::
109 * Machine Floating-Point Info Record::
110 * Multiple Response Sets Records::
111 * Extra Product Info Record::
112 * Variable Display Parameter Record::
113 * Long Variable Names Record::
114 * Very Long String Record::
115 * Character Encoding Record::
116 * Long String Value Labels Record::
117 * Long String Missing Values Record::
118 * Data File and Variable Attributes Records::
119 * Extended Number of Cases Record::
120 * Miscellaneous Informational Records::
121 * Dictionary Termination Record::
123 * Encrypted System Files::
126 @node File Header Record
127 @section File Header Record
129 The file header is always the first record in the file. It has the
136 int32 nominal_case_size;
141 char creation_date[9];
142 char creation_time[8];
148 @item char rec_type[4];
149 Record type code, either @samp{$FL2} for system files with
150 uncompressed data or data compressed with simple bytecode compression,
151 or @samp{$FL3} for system files with ZLIB compressed data.
153 This is truly a character field that uses the character encoding as
154 other strings. Thus, in a file with an ASCII-based character encoding
155 this field contains @code{24 46 4c 32} or @code{24 46 4c 33}, and in a
156 file with an EBCDIC-based encoding this field contains @code{5b c6 d3
157 f2}. (No EBCDIC-based ZLIB-compressed files have been observed.)
159 @item char prod_name[60];
160 Product identification string. This always begins with the characters
161 @samp{@@(#) SPSS DATA FILE}. PSPP uses the remaining characters to
162 give its version and the operating system name; for example, @samp{GNU
163 pspp 0.1.4 - sparc-sun-solaris2.5.2}. The string is truncated if it
164 would be longer than 60 characters; otherwise it is padded on the right
168 @item int32 layout_code;
169 Normally set to 2, although a few system files have been spotted in
170 the wild with a value of 3 here. PSPP use this value to determine the
171 file's integer endianness (@pxref{System File Format}).
173 @item int32 nominal_case_size;
174 Number of data elements per case. This is the number of variables,
175 except that long string variables add extra data elements (one for every
176 8 characters after the first 8). However, string variables do not
177 contribute to this value beyond the first 255 bytes. Further, system
178 files written by some systems set this value to -1. In general, it is
179 unsafe for systems reading system files to rely upon this value.
181 @item int32 compressed;
182 Set to 0 if the data in the file is not compressed, 1 if the data is
183 compressed with simple bytecode compression, 2 if the data is ZLIB
184 compressed. This field has value 2 if and only if @code{rec_type} is
187 @item int32 weight_index;
188 If one of the variables in the data set is used as a weighting
189 variable, set to the dictionary index of that variable, plus 1
190 (@pxref{Dictionary Index}). Otherwise, set to 0.
193 Set to the number of cases in the file if it is known, or -1 otherwise.
195 In the general case it is not possible to determine the number of cases
196 that will be output to a system file at the time that the header is
197 written. The way that this is dealt with is by writing the entire
198 system file, including the header, then seeking back to the beginning of
199 the file and writing just the @code{ncases} field. For files in which
200 this is not valid, the seek operation fails. In this case,
201 @code{ncases} remains -1.
205 Compression bias, ordinarily set to 100. Only integers between
206 @code{1 - bias} and @code{251 - bias} can be compressed.
208 By assuming that its value is 100, PSPP uses @code{bias} to determine
209 the file's floating-point format and endianness (@pxref{System File
210 Format}). If the compression bias is not 100, PSPP cannot auto-detect
211 the floating-point format and assumes that it is IEEE 754 format with
212 the same endianness as the system file's integers, which is correct
213 for all known system files.
215 @item char creation_date[9];
216 Date of creation of the system file, in @samp{dd mmm yy}
217 format, with the month as standard English abbreviations, using an
218 initial capital letter and following with lowercase. If the date is not
219 available then this field is arbitrarily set to @samp{01 Jan 70}.
221 @item char creation_time[8];
222 Time of creation of the system file, in @samp{hh:mm:ss}
223 format and using 24-hour time. If the time is not available then this
224 field is arbitrarily set to @samp{00:00:00}.
226 @item char file_label[64];
227 File label declared by the user, if any (@pxref{FILE LABEL,,,pspp,
228 PSPP Users Guide}). Padded on the right with spaces.
230 A product that identifies itself as @code{VOXCO INTERVIEWER 4.3} uses
231 CR-only line ends in this field, rather than the more usual LF-only or
234 @item char padding[3];
235 Ignored padding bytes to make the structure a multiple of 32 bits in
236 length. Set to zeros.
239 @node Variable Record
240 @section Variable Record
242 There must be one variable record for each numeric variable and each
243 string variable with width 8 bytes or less. String variables wider
244 than 8 bytes have one variable record for each 8 bytes, rounding up.
245 The first variable record for a long string specifies the variable's
246 correct dictionary information. Subsequent variable records for a
247 long string are filled with dummy information: a type of -1, no
248 variable label or missing values, print and write formats that are
249 ignored, and an empty string as name. A few system files have been
250 encountered that include a variable label on dummy variable records,
251 so readers should take care to parse dummy variable records in the
252 same way as other variable records.
254 @anchor{Dictionary Index}
255 The @dfn{dictionary index} of a variable is its offset in the set of
256 variable records, including dummy variable records for long string
257 variables. The first variable record has a dictionary index of 0, the
258 second has a dictionary index of 1, and so on.
260 The system file format does not directly support string variables
261 wider than 255 bytes. Such very long string variables are represented
262 by a number of narrower string variables. @xref{Very Long String
263 Record}, for details.
269 int32 n_missing_values;
274 /* @r{Present only if @code{has_var_label} is 1.} */
278 /* @r{Present only if @code{n_missing_values} is nonzero}. */
279 flt64 missing_values[];
283 @item int32 rec_type;
284 Record type code. Always set to 2.
287 Variable type code. Set to 0 for a numeric variable. For a short
288 string variable or the first part of a long string variable, this is set
289 to the width of the string. For the second and subsequent parts of a
290 long string variable, set to -1, and the remaining fields in the
291 structure are ignored.
293 @item int32 has_var_label;
294 If this variable has a variable label, set to 1; otherwise, set to 0.
296 @item int32 n_missing_values;
297 If the variable has no missing values, set to 0. If the variable has
298 one, two, or three discrete missing values, set to 1, 2, or 3,
299 respectively. If the variable has a range for missing variables, set to
300 -2; if the variable has a range for missing variables plus a single
301 discrete value, set to -3.
303 A long string variable always has the value 0 here. A separate record
304 indicates missing values for long string variables (@pxref{Long String
305 Missing Values Record}).
308 Print format for this variable. See below.
311 Write format for this variable. See below.
314 Variable name. The variable name must begin with a capital letter or
315 the at-sign (@samp{@@}). Subsequent characters may also be digits, octothorpes
316 (@samp{#}), dollar signs (@samp{$}), underscores (@samp{_}), or full
317 stops (@samp{.}). The variable name is padded on the right with spaces.
319 The @samp{name} fields should be unique within a system file. System
320 files written by SPSS that contain very long string variables with
321 similar names sometimes contain duplicate names that are later
322 eliminated by resolving the very long string names (@pxref{Very Long
323 String Record}). PSPP handles duplicates by assigning them new,
326 @item int32 label_len;
327 This field is present only if @code{has_var_label} is set to 1. It is
328 set to the length, in characters, of the variable label. The
329 documented maximum length varies from 120 to 255 based on SPSS
330 version, but some files have been seen with longer labels. PSPP
331 accepts longer labels and truncates them to 255 bytes on input.
334 This field is present only if @code{has_var_label} is set to 1. It has
335 length @code{label_len}, rounded up to the nearest multiple of 32 bits.
336 The first @code{label_len} characters are the variable's variable label.
338 @item flt64 missing_values[];
339 This field is present only if @code{n_missing_values} is nonzero. It
340 has the same number of 8-byte elements as the absolute value of
341 @code{n_missing_values}. Each element is interpreted as a number for
342 numeric variables (with HIGHEST and LOWEST indicated as described in
343 the chapter introduction). For string variables of width less than 8
344 bytes, elements are right-padded with spaces; for string variables
345 wider than 8 bytes, only the first 8 bytes of each missing value are
346 specified, with the remainder implicitly all spaces.
348 For discrete missing values, each element represents one missing
349 value. When a range is present, the first element denotes the minimum
350 value in the range, and the second element denotes the maximum value
351 in the range. When a range plus a value are present, the third
352 element denotes the additional discrete missing value.
355 The @code{print} and @code{write} members of sysfile_variable are output
356 formats coded into @code{int32} types. The least-significant byte
357 of the @code{int32} represents the number of decimal places, and the
358 next two bytes in order of increasing significance represent field width
359 and format type, respectively. The most-significant byte is not
360 used and should be set to zero.
362 Format types are defined as follows:
365 @multitable {Value} {@code{DATETIME}}
451 A few system files have been observed in the wild with invalid
452 @code{write} fields, in particular with value 0. Readers should
453 probably treat invalid @code{print} or @code{write} fields as some
456 @node Value Labels Records
457 @section Value Labels Records
459 The value label records documented in this section are used for
460 numeric and short string variables only. Long string variables may
461 have value labels, but their value labels are recorded using a
462 different record type (@pxref{Long String Value Labels Record}).
464 The value label record has the following format:
470 /* @r{Repeated @code{label_cnt} times}. */
477 @item int32 rec_type;
478 Record type. Always set to 3.
480 @item int32 label_count;
481 Number of value labels present in this record.
484 The remaining fields are repeated @code{count} times. Each
485 repetition specifies one value label.
489 A numeric value or a short string value padded as necessary to 8 bytes
490 in length. Its type and width cannot be determined until the
491 following value label variables record (see below) is read.
493 @item char label_len;
494 The label's length, in bytes. The documented maximum length varies
495 from 60 to 120 based on SPSS version. PSPP supports value labels up
499 @code{label_len} bytes of the actual label, followed by up to 7 bytes
500 of padding to bring @code{label} and @code{label_len} together to a
501 multiple of 8 bytes in length.
504 The value label record is always immediately followed by a value label
505 variables record with the following format:
514 @item int32 rec_type;
515 Record type. Always set to 4.
517 @item int32 var_count;
518 Number of variables that the associated value labels from the value
519 label record are to be applied.
522 A list of dictionary indexes of variables to which to apply the value
523 labels (@pxref{Dictionary Index}). There are @code{var_count}
526 String variables wider than 8 bytes may not be specified in this list.
529 @node Document Record
530 @section Document Record
532 The document record, if present, has the following format:
541 @item int32 rec_type;
542 Record type. Always set to 6.
545 Number of lines of documents present.
547 @item char lines[][80];
548 Document lines. The number of elements is defined by @code{n_lines}.
549 Lines shorter than 80 characters are padded on the right with spaces.
552 @node Machine Integer Info Record
553 @section Machine Integer Info Record
555 The integer info record, if present, has the following format:
567 int32 version_revision;
569 int32 floating_point_rep;
570 int32 compression_code;
572 int32 character_code;
576 @item int32 rec_type;
577 Record type. Always set to 7.
580 Record subtype. Always set to 3.
583 Size of each piece of data in the data part, in bytes. Always set to 4.
586 Number of pieces of data in the data part. Always set to 8.
588 @item int32 version_major;
589 PSPP major version number. In version @var{x}.@var{y}.@var{z}, this
592 @item int32 version_minor;
593 PSPP minor version number. In version @var{x}.@var{y}.@var{z}, this
596 @item int32 version_revision;
597 PSPP version revision number. In version @var{x}.@var{y}.@var{z},
600 @item int32 machine_code;
601 Machine code. PSPP always set this field to value to -1, but other
604 @item int32 floating_point_rep;
605 Floating point representation code. For IEEE 754 systems this is 1.
606 IBM 370 sets this to 2, and DEC VAX E to 3.
608 @item int32 compression_code;
609 Compression code. Always set to 1, regardless of whether or how the
612 @item int32 endianness;
613 Machine endianness. 1 indicates big-endian, 2 indicates little-endian.
615 @item int32 character_code;
616 @anchor{character-code} Character code. The following values have
617 been actually observed in system files:
627 The @code{windows-1250} code page for Central European and Eastern
631 The @code{windows-1252} code page for Western European languages.
640 The following additional values are known to be defined:
650 Other Windows code page numbers are known to be generally valid.
652 Old versions of SPSS for Unix and Windows always wrote value 2 in this
653 field, regardless of the encoding in use. Newer versions also write
654 the character encoding as a string (see @ref{Character Encoding
658 @node Machine Floating-Point Info Record
659 @section Machine Floating-Point Info Record
661 The floating-point info record, if present, has the following format:
677 @item int32 rec_type;
678 Record type. Always set to 7.
681 Record subtype. Always set to 4.
684 Size of each piece of data in the data part, in bytes. Always set to 8.
687 Number of pieces of data in the data part. Always set to 3.
690 The system missing value.
693 The value used for HIGHEST in missing values.
696 The value used for LOWEST in missing values.
699 @node Multiple Response Sets Records
700 @section Multiple Response Sets Records
702 The system file format has two different types of records that
703 represent multiple response sets (@pxref{MRSETS,,,pspp, PSPP Users
704 Guide}). The first type of record describes multiple response sets
705 that can be understood by SPSS before version 14. The second type of
706 record, with a closely related format, is used for multiple dichotomy
707 sets that use the CATEGORYLABELS=COUNTEDVALUES feature added in
717 /* @r{Exactly @code{count} bytes of data.} */
722 @item int32 rec_type;
723 Record type. Always set to 7.
726 Record subtype. Set to 7 for records that describe multiple response
727 sets understood by SPSS before version 14, or to 19 for records that
728 describe dichotomy sets that use the CATEGORYLABELS=COUNTEDVALUES
729 feature added in version 14.
732 The size of each element in the @code{mrsets} member. Always set to 1.
735 The total number of bytes in @code{mrsets}.
738 Zero or more line feeds (byte 0x0a), followed by a series of multiple
739 response sets, each of which consists of the following:
743 The set's name (an identifier that begins with @samp{$}), in mixed
744 upper and lower case.
747 An equals sign (@samp{=}).
750 @samp{C} for a multiple category set, @samp{D} for a multiple
751 dichotomy set with CATEGORYLABELS=VARLABELS, or @samp{E} for a
752 multiple dichotomy set with CATEGORYLABELS=COUNTEDVALUES.
755 For a multiple dichotomy set with CATEGORYLABELS=COUNTEDVALUES, a
756 space, followed by a number expressed as decimal digits, followed by a
757 space. If LABELSOURCE=VARLABEL was specified on MRSETS, then the
758 number is 11; otherwise it is 1.@footnote{This part of the format may
759 not be fully understood, because only a single example of each
760 possibility has been examined.}
763 For either kind of multiple dichotomy set, the counted value, as a
764 positive integer count specified as decimal digits, followed by a
765 space, followed by as many string bytes as specified in the count. If
766 the set contains numeric variables, the string consists of the counted
767 integer value expressed as decimal digits. If the set contains string
768 variables, the string contains the counted string value. Either way,
769 the string may be padded on the right with spaces (older versions of
770 SPSS seem to always pad to a width of 8 bytes; newer versions don't).
776 The multiple response set's label, using the same format as for the
777 counted value for multiple dichotomy sets. A string of length 0 means
778 that the set does not have a label. A string of length 0 is also
779 written if LABELSOURCE=VARLABEL was specified.
785 The short names of the variables in the set, converted to lowercase,
786 each separated from the previous by a single space.
789 One line feed (byte 0x0a). Sometimes multiple, even hundreds, of line
794 Example: Given appropriate variable definitions, consider the
795 following MRSETS command:
798 MRSETS /MCGROUP NAME=$a LABEL='my mcgroup' VARIABLES=a b c
799 /MDGROUP NAME=$b VARIABLES=g e f d VALUE=55
800 /MDGROUP NAME=$c LABEL='mdgroup #2' VARIABLES=h i j VALUE='Yes'
801 /MDGROUP NAME=$d LABEL='third mdgroup' CATEGORYLABELS=COUNTEDVALUES
802 VARIABLES=k l m VALUE=34
803 /MDGROUP NAME=$e CATEGORYLABELS=COUNTEDVALUES LABELSOURCE=VARLABEL
804 VARIABLES=n o p VALUE='choice'.
807 The above would generate the following multiple response set record of
811 $a=C 10 my mcgroup a b c
813 $c=D3 Yes 10 mdgroup #2 h i j
816 It would also generate the following multiple response set record with
820 $d=E 1 2 34 13 third mdgroup k l m
821 $e=E 11 6 choice 0 n o p
824 @node Extra Product Info Record
825 @section Extra Product Info Record
827 This optional record appears to contain a text string that describes
828 the program that wrote the file and the source of the data. (This is
829 redundant with the file label and product info found in the file
839 /* @r{Exactly @code{count} bytes of data.} */
844 @item int32 rec_type;
845 Record type. Always set to 7.
848 Record subtype. Always set to 10.
851 The size of each element in the @code{info} member. Always set to 1.
854 The total number of bytes in @code{info}.
857 A text string. A product that identifies itself as @code{VOXCO
858 INTERVIEWER 4.3} uses CR-only line ends in this field, rather than the
859 more usual LF-only or CR LF line ends.
862 @node Variable Display Parameter Record
863 @section Variable Display Parameter Record
865 The variable display parameter record, if present, has the following
875 /* @r{Repeated @code{count} times}. */
877 int32 width; /* @r{Not always present.} */
882 @item int32 rec_type;
883 Record type. Always set to 7.
886 Record subtype. Always set to 11.
889 The size of @code{int32}. Always set to 4.
892 The number of sets of variable display parameters (ordinarily the
893 number of variables in the dictionary), times 2 or 3.
896 The remaining members are repeated @code{count} times, in the same
897 order as the variable records. No element corresponds to variable
898 records that continue long string variables. The meanings of these
899 members are as follows:
903 The measurement type of the variable:
913 SPSS sometimes writes a @code{measure} of 0. PSPP interprets this as
917 The width of the display column for the variable in characters.
919 This field is present if @var{count} is 3 times the number of
920 variables in the dictionary. It is omitted if @var{count} is 2 times
921 the number of variables.
923 @item int32 alignment;
924 The alignment of the variable for display purposes:
936 @node Long Variable Names Record
937 @section Long Variable Names Record
939 If present, the long variable names record has the following format:
948 /* @r{Exactly @code{count} bytes of data.} */
949 char var_name_pairs[];
953 @item int32 rec_type;
954 Record type. Always set to 7.
957 Record subtype. Always set to 13.
960 The size of each element in the @code{var_name_pairs} member. Always set to 1.
963 The total number of bytes in @code{var_name_pairs}.
965 @item char var_name_pairs[];
966 A list of @var{key}--@var{value} tuples, where @var{key} is the name
967 of a variable, and @var{value} is its long variable name.
968 The @var{key} field is at most 8 bytes long and must match the
969 name of a variable which appears in the variable record (@pxref{Variable
971 The @var{value} field is at most 64 bytes long.
972 The @var{key} and @var{value} fields are separated by a @samp{=} byte.
973 Each tuple is separated by a byte whose value is 09. There is no
974 trailing separator following the last tuple.
975 The total length is @code{count} bytes.
978 @node Very Long String Record
979 @section Very Long String Record
981 Old versions of SPSS limited string variables to a width of 255 bytes.
982 For backward compatibility with these older versions, the system file
983 format represents a string longer than 255 bytes, called a @dfn{very
984 long string}, as a collection of strings no longer than 255 bytes
985 each. The strings concatenated to make a very long string are called
986 its @dfn{segments}; for consistency, variables other than very long
987 strings are considered to have a single segment.
989 A very long string with a width of @var{w} has @var{n} =
990 (@var{w} + 251) / 252 segments, that is, one segment for every
991 252 bytes of width, rounding up. It would be logical, then, for each
992 of the segments except the last to have a width of 252 and the last
993 segment to have the remainder, but this is not the case. In fact,
994 each segment except the last has a width of 255 bytes. The last
995 segment has width @var{w} - (@var{n} - 1) * 252; some versions
996 of SPSS make it slightly wider, but not wide enough to make the last
997 segment require another 8 bytes of data.
999 Data is packed tightly into segments of a very long string, 255 bytes
1000 per segment. Because 255 bytes of segment data are allocated for
1001 every 252 bytes of the very long string's width (approximately), some
1002 unused space is left over at the end of the allocated segments. Data
1003 in unused space is ignored.
1005 Example: Consider a very long string of width 20,000. Such a very
1006 long string has 20,000 / 252 = 80 (rounding up) segments. The first
1007 79 segments have width 255; the last segment has width 20,000 - 79 *
1008 252 = 92 or slightly wider (up to 96 bytes, the next multiple of 8).
1009 The very long string's data is actually stored in the 19,890 bytes in
1010 the first 78 segments, plus the first 110 bytes of the 79th segment
1011 (19,890 + 110 = 20,000). The remaining 145 bytes of the 79th segment
1012 and all 92 bytes of the 80th segment are unused.
1014 The very long string record explains how to stitch together segments
1015 to obtain very long string data. For each of the very long string
1016 variables in the dictionary, it specifies the name of its first
1017 segment's variable and the very long string variable's actual width.
1018 The remaining segments immediately follow the named variable in the
1019 system file's dictionary.
1021 The very long string record, which is present only if the system file
1022 contains very long string variables, has the following format:
1031 /* @r{Exactly @code{count} bytes of data.} */
1032 char string_lengths[];
1036 @item int32 rec_type;
1037 Record type. Always set to 7.
1039 @item int32 subtype;
1040 Record subtype. Always set to 14.
1043 The size of each element in the @code{string_lengths} member. Always set to 1.
1046 The total number of bytes in @code{string_lengths}.
1048 @item char string_lengths[];
1049 A list of @var{key}--@var{value} tuples, where @var{key} is the name
1050 of a variable, and @var{value} is its length.
1051 The @var{key} field is at most 8 bytes long and must match the
1052 name of a variable which appears in the variable record (@pxref{Variable
1054 The @var{value} field is exactly 5 bytes long. It is a zero-padded,
1055 ASCII-encoded string that is the length of the variable.
1056 The @var{key} and @var{value} fields are separated by a @samp{=} byte.
1057 Tuples are delimited by a two-byte sequence @{00, 09@}.
1058 After the last tuple, there may be a single byte 00, or @{00, 09@}.
1059 The total length is @code{count} bytes.
1062 @node Character Encoding Record
1063 @section Character Encoding Record
1065 This record, if present, indicates the character encoding for string data,
1066 long variable names, variable labels, value labels and other strings in the
1076 /* @r{Exactly @code{count} bytes of data.} */
1081 @item int32 rec_type;
1082 Record type. Always set to 7.
1084 @item int32 subtype;
1085 Record subtype. Always set to 20.
1088 The size of each element in the @code{encoding} member. Always set to 1.
1091 The total number of bytes in @code{encoding}.
1093 @item char encoding[];
1094 The name of the character encoding. Normally this will be an official
1095 IANA character set name or alias.
1096 See @url{http://www.iana.org/assignments/character-sets}.
1097 Character set names are not case-sensitive, but SPSS appears to write
1098 them in all-uppercase.
1101 This record is not present in files generated by older software. See
1102 also the @code{character_code} field in the machine integer info
1103 record (@pxref{character-code}).
1105 When the character encoding record and the machine integer info record
1106 are both present, all system files observed in practice indicate the
1107 same character encoding, e.g.@: 1252 as @code{character_code} and
1108 @code{windows-1252} as @code{encoding}, 65001 and @code{UTF-8}, etc.
1110 If, for testing purposes, a file is crafted with different
1111 @code{character_code} and @code{encoding}, it seems that
1112 @code{character_code} controls the encoding for all strings in the
1113 system file before the dictionary termination record, including
1114 strings in data (e.g.@: string missing values), and @code{encoding}
1115 controls the encoding for strings following the dictionary termination
1118 @node Long String Value Labels Record
1119 @section Long String Value Labels Record
1121 This record, if present, specifies value labels for long string
1131 /* @r{Repeated up to exactly @code{count} bytes.} */
1136 long_string_label labels[];
1140 @item int32 rec_type;
1141 Record type. Always set to 7.
1143 @item int32 subtype;
1144 Record subtype. Always set to 21.
1150 The number of bytes following the header until the next header.
1152 @item int32 var_name_len;
1153 @itemx char var_name[];
1154 The number of bytes in the name of the variable that has long string
1155 value labels, plus the variable name itself, which consists of exactly
1156 @code{var_name_len} bytes. The variable name is not padded to any
1157 particular boundary, nor is it null-terminated.
1159 @item int32 var_width;
1160 The width of the variable, in bytes, which will be between 9 and
1163 @item int32 n_labels;
1164 @itemx long_string_label labels[];
1165 The long string labels themselves. The @code{labels} array contains
1166 exactly @code{n_labels} elements, each of which has the following
1177 @item int32 value_len;
1178 @itemx char value[];
1179 The string value being labeled. @code{value_len} is the number of
1180 bytes in @code{value}; it is equal to @code{var_width}. The
1181 @code{value} array is not padded or null-terminated.
1183 @item int32 label_len;
1184 @itemx char label[];
1185 The label for the string value. @code{label_len}, which must be
1186 between 0 and 120, is the number of bytes in @code{label}. The
1187 @code{label} array is not padded or null-terminated.
1191 @node Long String Missing Values Record
1192 @section Long String Missing Values Record
1194 This record, if present, specifies missing values for long string
1204 /* @r{Repeated up to exactly @code{count} bytes.} */
1207 char n_missing_values;
1208 long_string_missing_value values[];
1212 @item int32 rec_type;
1213 Record type. Always set to 7.
1215 @item int32 subtype;
1216 Record subtype. Always set to 22.
1222 The number of bytes following the header until the next header.
1224 @item int32 var_name_len;
1225 @itemx char var_name[];
1226 The number of bytes in the name of the long string variable that has
1227 missing values, plus the variable name itself, which consists of
1228 exactly @code{var_name_len} bytes. The variable name is not padded to
1229 any particular boundary, nor is it null-terminated.
1231 @item char n_missing_values;
1232 The number of missing values, either 1, 2, or 3. (This is, unusually,
1233 a single byte instead of a 32-bit number.)
1235 @item long_string_missing_value values[];
1236 The missing values themselves. This array contains exactly
1237 @code{n_missing_values} elements, each of which has the following
1246 @item int32 value_len;
1247 The length of the missing value string, in bytes. This value should
1248 be 8, because long string variables are at least 8 bytes wide (by
1249 definition), only the first 8 bytes of a long string variable's
1250 missing values are allowed to be non-spaces, and any spaces within the
1251 first 8 bytes are included in the missing value here.
1254 The missing value string, exactly @code{value_len} bytes, without
1255 any padding or null terminator.
1259 @node Data File and Variable Attributes Records
1260 @section Data File and Variable Attributes Records
1262 The data file and variable attributes records represent custom
1263 attributes for the system file or for individual variables in the
1264 system file, as defined on the DATAFILE ATTRIBUTE (@pxref{DATAFILE
1265 ATTRIBUTE,,,pspp, PSPP Users Guide}) and VARIABLE ATTRIBUTE commands
1266 (@pxref{VARIABLE ATTRIBUTE,,,pspp, PSPP Users Guide}), respectively.
1275 /* @r{Exactly @code{count} bytes of data.} */
1280 @item int32 rec_type;
1281 Record type. Always set to 7.
1283 @item int32 subtype;
1284 Record subtype. Always set to 17 for a data file attribute record or
1285 to 18 for a variable attributes record.
1288 The size of each element in the @code{attributes} member. Always set to 1.
1291 The total number of bytes in @code{attributes}.
1293 @item char attributes[];
1294 The attributes, in a text-based format.
1296 In record type 17, this field contains a single attribute set. An
1297 attribute set is a sequence of one or more attributes concatenated
1298 together. Each attribute consists of a name, which has the same
1299 syntax as a variable name, followed by, inside parentheses, a sequence
1300 of one or more values. Each value consists of a string enclosed in
1301 single quotes (@code{'}) followed by a line feed (byte 0x0a). A value
1302 may contain single quote characters, which are not themselves escaped
1303 or quoted or required to be present in pairs. There is no apparent
1304 way to embed a line feed in a value. There is no distinction between
1305 an attribute with a single value and an attribute array with one
1308 In record type 18, this field contains a sequence of one or more
1309 variable attribute sets. If more than one variable attribute set is
1310 present, each one after the first is delimited from the previous by
1311 @code{/}. Each variable attribute set consists of a long
1313 followed by @code{:}, followed by an attribute set with the same
1314 syntax as on record type 17.
1316 The total length is @code{count} bytes.
1321 A system file produced with the following VARIABLE ATTRIBUTE commands
1325 VARIABLE ATTRIBUTE VARIABLES=dummy ATTRIBUTE=fred[1]('23') fred[2]('34').
1326 VARIABLE ATTRIBUTE VARIABLES=dummy ATTRIBUTE=bert('123').
1330 will contain a variable attribute record with the following contents:
1333 0000 07 00 00 00 12 00 00 00 01 00 00 00 22 00 00 00 |............"...|
1334 0010 64 75 6d 6d 79 3a 66 72 65 64 28 27 32 33 27 0a |dummy:fred('23'.|
1335 0020 27 33 34 27 0a 29 62 65 72 74 28 27 31 32 33 27 |'34'.)bert('123'|
1343 @node Variable Roles
1344 @subsection Variable Roles
1346 A variable's role is represented as an attribute named @code{$@@Role}.
1347 This attribute has a single element whose values and their meanings
1352 Input. This, the default, is the most common role.
1365 @node Extended Number of Cases Record
1366 @section Extended Number of Cases Record
1368 The file header record expresses the number of cases in the system
1369 file as an int32 (@pxref{File Header Record}). This record allows the
1370 number of cases in the system file to be expressed as a 64-bit number.
1382 @item int32 rec_type;
1383 Record type. Always set to 7.
1385 @item int32 subtype;
1386 Record subtype. Always set to 16.
1389 Size of each element. Always set to 8.
1392 Number of pieces of data in the data part. Alway set to 2.
1394 @item int64 unknown;
1395 Meaning unknown. Always set to 1.
1397 @item int64 ncases64;
1398 Number of cases in the file as a 64-bit integer. Presumably this
1399 could be -1 to indicate that the number of cases is unknown, for the
1400 same reason as @code{ncases} in the file header record, but this has
1401 not been observed in the wild.
1404 @node Miscellaneous Informational Records
1405 @section Miscellaneous Informational Records
1407 Some specific types of miscellaneous informational records are
1408 documented here, but others are known to exist. PSPP ignores unknown
1409 miscellaneous informational records when reading system files.
1418 /* @r{Exactly @code{size * count} bytes of data.} */
1423 @item int32 rec_type;
1424 Record type. Always set to 7.
1426 @item int32 subtype;
1427 Record subtype. May take any value. According to Aapi
1428 H@"am@"al@"ainen, value 5 indicates a set of grouped variables and 6
1429 indicates date info (probably related to USE). Subtype 24 appears to
1430 contain XML that describes how data in the file should be displayed
1434 Size of each piece of data in the data part. Should have the value 1,
1435 4, or 8, for @code{char}, @code{int32}, and @code{flt64} format data,
1439 Number of pieces of data in the data part.
1442 Arbitrary data. There must be @code{size} times @code{count} bytes of
1446 @node Dictionary Termination Record
1447 @section Dictionary Termination Record
1449 The dictionary termination record separates all other records from the
1458 @item int32 rec_type;
1459 Record type. Always set to 999.
1462 Ignored padding. Should be set to 0.
1466 @section Data Record
1468 The data record must follow all other records in the system file.
1469 Every system file must have a data record that specifies data for at
1470 least one case. The format of the data record varies depending on the
1471 value of @code{compression} in the file header record:
1474 @item 0: no compression
1475 Data is arranged as a series of 8-byte elements.
1476 Each element corresponds to
1477 the variable declared in the respective variable record (@pxref{Variable
1478 Record}). Numeric values are given in @code{flt64} format; string
1479 values are literal characters string, padded on the right when
1480 necessary to fill out 8-byte units.
1482 @item 1: bytecode compression
1484 of the data record is divided into a series of 1-byte command
1485 codes. These codes have meanings as described below:
1489 Ignored. If the program writing the system file accumulates compressed
1490 data in blocks of fixed length, 0 bytes can be used to pad out extra
1491 bytes remaining at the end of a fixed-size block.
1495 value @var{code} - @var{bias}, where
1496 @var{code} is the value of the compression code and @var{bias} is the
1497 variable @code{bias} from the file header. For example,
1498 code 105 with bias 100.0 (the normal value) indicates a numeric variable
1500 One file has been seen written by SPSS 14 that contained such a code
1501 in a @emph{string} field with the value 0 (after the bias is
1502 subtracted) as a way of encoding null bytes.
1505 End of file. This code may or may not appear at the end of the data
1506 stream. PSPP always outputs this code but its use is not required.
1509 A numeric or string value that is not
1510 compressible. The value is stored in the 8 bytes following the
1511 current block of command bytes. If this value appears twice in a block
1512 of command bytes, then it indicates the second group of 8 bytes following the
1513 command bytes, and so on.
1516 An 8-byte string value that is all spaces.
1519 The system-missing value.
1522 The end of the 8-byte group of bytecodes is followed by any 8-byte
1523 blocks of non-compressible values indicated by code 253. After that
1524 follows another 8-byte group of bytecodes, then those bytecodes'
1525 non-compressible values. The pattern repeats to the end of the file
1526 or a code with value 252.
1528 @item 2: ZLIB compression
1529 The data record consists of the following, in order:
1533 ZLIB data header, 24 bytes long.
1536 One or more variable-length blocks of ZLIB compressed data.
1539 ZLIB data trailer, with a 24-byte fixed header plus an additional 24
1540 bytes for each preceding ZLIB compressed data block.
1543 The ZLIB data header has the following format:
1552 @item int64 zheader_ofs;
1553 The offset, in bytes, of the beginning of this structure within the
1556 @item int64 ztrailer_ofs;
1557 The offset, in bytes, of the first byte of the ZLIB data trailer.
1559 @item int64 ztrailer_len;
1560 The number of bytes in the ZLIB data trailer. This and the previous
1561 field sum to the size of the system file in bytes.
1564 The data header is followed by @code{(ztrailer_ofs - 24) / 24} ZLIB
1565 compressed data blocks. Each ZLIB compressed data block begins with a
1566 ZLIB header as specified in RFC@tie{}1950, e.g.@: hex bytes @code{78
1567 01} (the only header yet observed in practice). Each block
1568 decompresses to a fixed number of bytes (in practice only
1569 @code{0x3ff000}-byte blocks have been observed), except that the last
1570 block of data may be shorter. The last ZLIB compressed data block
1571 gends just before offset @code{ztrailer_ofs}.
1573 The result of ZLIB decompression is bytecode compressed data as
1574 described above for compression format 1.
1576 The ZLIB data trailer begins with the following 24-byte fixed header:
1586 @item int64 int_bias;
1587 The compression bias as a negative integer, e.g.@: if @code{bias} in
1588 the file header record is 100.0, then @code{int_bias} is @minus{}100
1589 (this is the only value yet observed in practice).
1592 Always observed to be zero.
1594 @item int32 block_size;
1595 The number of bytes in each ZLIB compressed data block, except
1596 possibly the last, following decompression. Only @code{0x3ff000} has
1597 been observed so far.
1599 @item int32 n_blocks;
1600 The number of ZLIB compressed data blocks, always exactly
1601 @code{(ztrailer_ofs - 24) / 24}.
1604 The fixed header is followed by @code{n_blocks} 24-byte ZLIB data
1605 block descriptors, each of which describes the compressed data block
1606 corresponding to its offset. Each block descriptor has the following
1610 int64 uncompressed_ofs;
1611 int64 compressed_ofs;
1612 int32 uncompressed_size;
1613 int32 compressed_size;
1617 @item int64 uncompressed_ofs;
1618 The offset, in bytes, that this block of data would have in a similar
1619 system file that uses compression format 1. This is
1620 @code{zheader_ofs} in the first block descriptor, and in each
1621 succeeding block descriptor it is the sum of the previous desciptor's
1622 @code{uncompressed_ofs} and @code{uncompressed_size}.
1624 @item int64 compressed_ofs;
1625 The offset, in bytes, of the actual beginning of this compressed data
1626 block. This is @code{zheader_ofs + 24} in the first block descriptor,
1627 and in each succeeding block descriptor it is the sum of the previous
1628 descriptor's @code{compressed_ofs} and @code{compressed_size}. The
1629 final block descriptor's @code{compressed_ofs} and
1630 @code{compressed_size} sum to @code{ztrailer_ofs}.
1632 @item int32 uncompressed_size;
1633 The number of bytes in this data block, after decompression. This is
1634 @code{block_size} in every data block except the last, which may be
1637 @item int32 compressed_size;
1638 The number of bytes in this data block, as stored compressed in this
1643 @setfilename ignored
1645 @node Encrypted System Files
1646 @section Encrypted System Files
1648 SPSS 21 and later support an encrypted system file format.
1651 The SPSS encrypted file format is poorly designed. It is much cheaper
1652 and faster to decrypt a file encrypted this way than if a well
1653 designed alternative were used. If you must use this format, use a
1654 10-byte randomly generated password.
1657 @subheading Encrypted File Format
1659 Encrypted system files begin with the following 36-byte fixed header:
1662 0000 1c 00 00 00 00 00 00 00 45 4e 43 52 59 50 54 45 |........ENCRYPTE|
1663 0010 44 53 41 56 15 00 00 00 00 00 00 00 00 00 00 00 |DSAV............|
1664 0020 00 00 00 00 |....|
1667 Following the fixed header is a complete system file in the usual
1668 format, except that each 16-byte block is encrypted with AES-256 in
1669 ECB mode. The AES-256 key is derived from a password in the following
1674 Start from the literal password typed by the user. Truncate it to at
1675 most 10 bytes, then append (between 1 and 22) null bytes until there
1676 are exactly 32 bytes. Call this @var{password}.
1679 Let @var{constant} be the following 73-byte constant:
1682 0000 00 00 00 01 35 27 13 cc 53 a7 78 89 87 53 22 11
1683 0010 d6 5b 31 58 dc fe 2e 7e 94 da 2f 00 cc 15 71 80
1684 0020 0a 6c 63 53 00 38 c3 38 ac 22 f3 63 62 0e ce 85
1685 0030 3f b8 07 4c 4e 2b 77 c7 21 f5 1a 80 1d 67 fb e1
1686 0040 e1 83 07 d8 0d 00 00 01 00
1690 Compute CMAC-AES-256(@var{password}, @var{constant}). Call the
1691 16-byte result @var{cmac}.
1694 The 32-byte AES-256 key is @var{cmac} || @var{cmac}, that is,
1695 @var{cmac} repeated twice.
1698 @subsubheading Example
1700 Consider the password @samp{pspp}. @var{password} is:
1703 0000 70 73 70 70 00 00 00 00 00 00 00 00 00 00 00 00 |pspp............|
1704 0010 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 |................|
1711 0000 3e da 09 8e 66 04 d4 fd f9 63 0c 2c a8 6f b0 45
1718 0000 3e da 09 8e 66 04 d4 fd f9 63 0c 2c a8 6f b0 45
1719 0010 3e da 09 8e 66 04 d4 fd f9 63 0c 2c a8 6f b0 45
1722 @subheading Password Encoding
1724 SPSS also supports what it calls ``encrypted passwords.'' These are
1725 not encrypted. They are encoded with a simple, fixed scheme. An
1726 encoded password is always a multiple of 2 characters long, and never
1727 longer than 20 characters. The characters in an encoded password are
1728 always in the graphic ASCII range 33 through 126. Each successive
1729 pair of characters in the password encodes a single byte in the
1732 Use the following algorithm to decode a pair of characters:
1736 Let @var{a} be the ASCII code of the first character, and @var{b} be
1737 the ASCII code of the second character.
1740 Let @var{ah} be the most significant 4 bits of @var{a}. Find the line
1741 in the table below that has @var{ah} on the left side. The right side
1742 of the line is a set of possible values for the most significant 4
1743 bits of the decoded byte.
1746 @t{2 } @result{} @t{2367}
1747 @t{3 } @result{} @t{0145}
1748 @t{47} @result{} @t{89cd}
1749 @t{56} @result{} @t{abef}
1753 Let @var{bh} be the most significant 4 bits of @var{b}. Find the line
1754 in the second table below that has @var{bh} on the left side. The
1755 right side of the line is a set of possible values for the most
1756 significant 4 bits of the decoded byte. Together with the results of
1757 the previous step, only a single possibility is left.
1760 @t{2 } @result{} @t{139b}
1761 @t{3 } @result{} @t{028a}
1762 @t{47} @result{} @t{46ce}
1763 @t{56} @result{} @t{57df}
1767 Let @var{al} be the least significant 4 bits of @var{a}. Find the
1768 line in the table below that has @var{al} on the left side. The right
1769 side of the line is a set of possible values for the least significant
1770 4 bits of the decoded byte.
1773 @t{03cf} @result{} @t{0145}
1774 @t{12de} @result{} @t{2367}
1775 @t{478b} @result{} @t{89cd}
1776 @t{569a} @result{} @t{abef}
1780 Let @var{bl} be the least significant 4 bits of @var{b}. Find the
1781 line in the table below that has @var{bl} on the left side. The right
1782 side of the line is a set of possible values for the least significant
1783 4 bits of the decoded byte. Together with the results of the previous
1784 step, only a single possibility is left.
1787 @t{03cf} @result{} @t{028a}
1788 @t{12de} @result{} @t{139b}
1789 @t{478b} @result{} @t{46ce}
1790 @t{569a} @result{} @t{57df}
1794 @subsubheading Example
1796 Consider the encoded character pair @samp{-|}. @var{a} is
1797 0x2d and @var{b} is 0x7c, so @var{ah} is 2, @var{bh} is 7, @var{al} is
1798 0xd, and @var{bl} is 0xc. @var{ah} means that the most significant
1799 four bits of the decoded character is 2, 3, 6, or 7, and @var{bh}
1800 means that they are 4, 6, 0xc, or 0xe. The single possibility in
1801 common is 6, so the most significant four bits are 6. Similarly,
1802 @var{al} means that the least significant four bits are 2, 3, 6, or 7,
1803 and @var{bl} means they are 0, 2, 8, or 0xa, so the least significant
1804 four bits are 2. The decoded character is therefore 0x62, the letter