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 have 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}).
72 * System File Record Structure::
73 * File Header Record::
75 * Value Labels Records::
77 * Machine Integer Info Record::
78 * Machine Floating-Point Info Record::
79 * Multiple Response Sets Records::
80 * Extra Product Info Record::
81 * Variable Display Parameter Record::
82 * Long Variable Names Record::
83 * Very Long String Record::
84 * Character Encoding Record::
85 * Long String Value Labels Record::
86 * Long String Missing Values Record::
87 * Data File and Variable Attributes Records::
88 * Extended Number of Cases Record::
89 * Other Informational Records::
90 * Dictionary Termination Record::
94 @node System File Record Structure
95 @section System File Record Structure
97 System files are divided into records with the following format:
104 This header does not identify the length of the @code{data} or any
105 information about what it contains, so the system file reader must
106 understand the format of @code{data} based on @code{type}. However,
107 records with type 7, called @dfn{extension records}, have a stricter
115 char data[size * count];
119 @item int32 rec_type;
120 Record type. Always set to 7.
123 Record subtype. This value identifies a particular kind of extension
127 The size of each piece of data that follows the header, in bytes.
128 Known extension records use 1, 4, or 8, for @code{char}, @code{int32},
129 and @code{flt64} format data, respectively.
132 The number of pieces of data that follow the header.
134 @item char data[size * count];
135 Data, whose format and interpretation depend on the subtype.
138 An extension record contains exactly @code{size * count} bytes of
139 data, which allows a reader that does not understand an extension
140 record to skip it. Extension records provide only nonessential
141 information, so this allows for files written by newer software to
142 preserve backward compatibility with older or less capable readers.
144 Records in a system file must appear in the following order:
154 All pairs of value labels records and value label variables records,
158 Document record, if present.
161 Extension (type 7) records, in ascending numerical order of their
164 System files written by SPSS include at most one of each kind of
165 extension record. This is generally true of system files written by
166 other software as well, with known exceptions noted below in the
167 individual sections about each type of record.
170 Dictionary termination record.
176 We advise authors of programs that read system files to tolerate
177 format variations. Various kinds of misformatting and corruption have
178 been observed in system files written by SPSS and other software
179 alike. In particular, because extension records provide nonessential
180 information, it is generally better to ignore an extension record
181 entirely than to refuse to read a system file.
183 The following sections describe the known kinds of records.
185 @node File Header Record
186 @section File Header Record
188 A system file begins with the file header, with the following format:
194 int32 nominal_case_size;
199 char creation_date[9];
200 char creation_time[8];
206 @item char rec_type[4];
207 Record type code, either @samp{$FL2} for system files with
208 uncompressed data or data compressed with simple bytecode compression,
209 or @samp{$FL3} for system files with ZLIB compressed data.
211 This is truly a character field that uses the character encoding as
212 other strings. Thus, in a file with an ASCII-based character encoding
213 this field contains @code{24 46 4c 32} or @code{24 46 4c 33}, and in a
214 file with an EBCDIC-based encoding this field contains @code{5b c6 d3
215 f2}. (No EBCDIC-based ZLIB-compressed files have been observed.)
217 @item char prod_name[60];
218 Product identification string. This always begins with the characters
219 @samp{@@(#) SPSS DATA FILE}. PSPP uses the remaining characters to
220 give its version and the operating system name; for example, @samp{GNU
221 pspp 0.1.4 - sparc-sun-solaris2.5.2}. The string is truncated if it
222 would be longer than 60 characters; otherwise it is padded on the right
226 @item int32 layout_code;
227 Normally set to 2, although a few system files have been spotted in
228 the wild with a value of 3 here. PSPP use this value to determine the
229 file's integer endianness (@pxref{System File Format}).
231 @item int32 nominal_case_size;
232 Number of data elements per case. This is the number of variables,
233 except that long string variables add extra data elements (one for every
234 8 characters after the first 8). However, string variables do not
235 contribute to this value beyond the first 255 bytes. Further, some
236 software always writes -1 or 0 in this field. In general, it is
237 unsafe for systems reading system files to rely upon this value.
239 @item int32 compression;
240 Set to 0 if the data in the file is not compressed, 1 if the data is
241 compressed with simple bytecode compression, 2 if the data is ZLIB
242 compressed. This field has value 2 if and only if @code{rec_type} is
245 @item int32 weight_index;
246 If one of the variables in the data set is used as a weighting
247 variable, set to the dictionary index of that variable, plus 1
248 (@pxref{Dictionary Index}). Otherwise, set to 0.
251 Set to the number of cases in the file if it is known, or -1 otherwise.
253 In the general case it is not possible to determine the number of cases
254 that will be output to a system file at the time that the header is
255 written. The way that this is dealt with is by writing the entire
256 system file, including the header, then seeking back to the beginning of
257 the file and writing just the @code{ncases} field. For files in which
258 this is not valid, the seek operation fails. In this case,
259 @code{ncases} remains -1.
263 Compression bias, ordinarily set to 100. Only integers between
264 @code{1 - bias} and @code{251 - bias} can be compressed.
266 By assuming that its value is 100, PSPP uses @code{bias} to determine
267 the file's floating-point format and endianness (@pxref{System File
268 Format}). If the compression bias is not 100, PSPP cannot auto-detect
269 the floating-point format and assumes that it is IEEE 754 format with
270 the same endianness as the system file's integers, which is correct
271 for all known system files.
273 @item char creation_date[9];
274 Date of creation of the system file, in @samp{dd mmm yy}
275 format, with the month as standard English abbreviations, using an
276 initial capital letter and following with lowercase. If the date is not
277 available then this field is arbitrarily set to @samp{01 Jan 70}.
279 @item char creation_time[8];
280 Time of creation of the system file, in @samp{hh:mm:ss}
281 format and using 24-hour time. If the time is not available then this
282 field is arbitrarily set to @samp{00:00:00}.
284 @item char file_label[64];
285 File label declared by the user, if any (@pxref{FILE LABEL,,,pspp,
286 PSPP Users Guide}). Padded on the right with spaces.
288 A product that identifies itself as @code{VOXCO INTERVIEWER 4.3} uses
289 CR-only line ends in this field, rather than the more usual LF-only or
292 @item char padding[3];
293 Ignored padding bytes to make the structure a multiple of 32 bits in
294 length. Set to zeros.
297 @node Variable Record
298 @section Variable Record
300 There must be one variable record for each numeric variable and each
301 string variable with width 8 bytes or less. String variables wider
302 than 8 bytes have one variable record for each 8 bytes, rounding up.
303 The first variable record for a long string specifies the variable's
304 correct dictionary information. Subsequent variable records for a
305 long string are filled with dummy information: a type of -1, no
306 variable label or missing values, print and write formats that are
307 ignored, and an empty string as name. A few system files have been
308 encountered that include a variable label on dummy variable records,
309 so readers should take care to parse dummy variable records in the
310 same way as other variable records.
312 @anchor{Dictionary Index}
313 The @dfn{dictionary index} of a variable is its offset in the set of
314 variable records, including dummy variable records for long string
315 variables. The first variable record has a dictionary index of 0, the
316 second has a dictionary index of 1, and so on.
318 The system file format does not directly support string variables
319 wider than 255 bytes. Such very long string variables are represented
320 by a number of narrower string variables. @xref{Very Long String
321 Record}, for details.
323 A system file should contain at least one variable and thus at least
324 one variable record, but system files have been observed in the wild
325 without any variables (thus, no data either).
331 int32 n_missing_values;
336 /* @r{Present only if @code{has_var_label} is 1.} */
340 /* @r{Present only if @code{n_missing_values} is nonzero}. */
341 flt64 missing_values[];
345 @item int32 rec_type;
346 Record type code. Always set to 2.
349 Variable type code. Set to 0 for a numeric variable. For a short
350 string variable or the first part of a long string variable, this is set
351 to the width of the string. For the second and subsequent parts of a
352 long string variable, set to -1, and the remaining fields in the
353 structure are ignored.
355 @item int32 has_var_label;
356 If this variable has a variable label, set to 1; otherwise, set to 0.
358 @item int32 n_missing_values;
359 If the variable has no missing values, set to 0. If the variable has
360 one, two, or three discrete missing values, set to 1, 2, or 3,
361 respectively. If the variable has a range for missing variables, set to
362 -2; if the variable has a range for missing variables plus a single
363 discrete value, set to -3.
365 A long string variable always has the value 0 here. A separate record
366 indicates missing values for long string variables (@pxref{Long String
367 Missing Values Record}).
370 Print format for this variable. See below.
373 Write format for this variable. See below.
376 Variable name. The variable name must begin with a capital letter or
377 the at-sign (@samp{@@}). Subsequent characters may also be digits, octothorpes
378 (@samp{#}), dollar signs (@samp{$}), underscores (@samp{_}), or full
379 stops (@samp{.}). The variable name is padded on the right with spaces.
381 The @samp{name} fields should be unique within a system file. System
382 files written by SPSS that contain very long string variables with
383 similar names sometimes contain duplicate names that are later
384 eliminated by resolving the very long string names (@pxref{Very Long
385 String Record}). PSPP handles duplicates by assigning them new,
388 @item int32 label_len;
389 This field is present only if @code{has_var_label} is set to 1. It is
390 set to the length, in characters, of the variable label. The
391 documented maximum length varies from 120 to 255 based on SPSS
392 version, but some files have been seen with longer labels. PSPP
393 accepts labels of any length.
396 This field is present only if @code{has_var_label} is set to 1. It has
397 length @code{label_len}, rounded up to the nearest multiple of 32 bits.
398 The first @code{label_len} characters are the variable's variable label.
400 @item flt64 missing_values[];
401 This field is present only if @code{n_missing_values} is nonzero. It
402 has the same number of 8-byte elements as the absolute value of
403 @code{n_missing_values}. Each element is interpreted as a number for
404 numeric variables (with HIGHEST and LOWEST indicated as described in
405 the chapter introduction). For string variables of width less than 8
406 bytes, elements are right-padded with spaces; for string variables
407 wider than 8 bytes, only the first 8 bytes of each missing value are
408 specified, with the remainder implicitly all spaces.
410 For discrete missing values, each element represents one missing
411 value. When a range is present, the first element denotes the minimum
412 value in the range, and the second element denotes the maximum value
413 in the range. When a range plus a value are present, the third
414 element denotes the additional discrete missing value.
417 @anchor{System File Output Formats}
418 The @code{print} and @code{write} members of sysfile_variable are output
419 formats coded into @code{int32} types. The least-significant byte
420 of the @code{int32} represents the number of decimal places, and the
421 next two bytes in order of increasing significance represent field width
422 and format type, respectively. The most-significant byte is not
423 used and should be set to zero.
425 Format types are defined as follows:
428 @multitable {Value} {@code{DATETIME}}
514 A few system files have been observed in the wild with invalid
515 @code{write} fields, in particular with value 0. Readers should
516 probably treat invalid @code{print} or @code{write} fields as some
519 @node Value Labels Records
520 @section Value Labels Records
522 The value label records documented in this section are used for
523 numeric and short string variables only. Long string variables may
524 have value labels, but their value labels are recorded using a
525 different record type (@pxref{Long String Value Labels Record}).
527 The value label record has the following format:
533 /* @r{Repeated @code{label_cnt} times}. */
540 @item int32 rec_type;
541 Record type. Always set to 3.
543 @item int32 label_count;
544 Number of value labels present in this record.
547 The remaining fields are repeated @code{count} times. Each
548 repetition specifies one value label.
552 A numeric value or a short string value padded as necessary to 8 bytes
553 in length. Its type and width cannot be determined until the
554 following value label variables record (see below) is read.
556 @item char label_len;
557 The label's length, in bytes. The documented maximum length varies
558 from 60 to 120 based on SPSS version. PSPP supports value labels up
562 @code{label_len} bytes of the actual label, followed by up to 7 bytes
563 of padding to bring @code{label} and @code{label_len} together to a
564 multiple of 8 bytes in length.
567 The value label record is always immediately followed by a value label
568 variables record with the following format:
577 @item int32 rec_type;
578 Record type. Always set to 4.
580 @item int32 var_count;
581 Number of variables that the associated value labels from the value
582 label record are to be applied.
585 A list of dictionary indexes of variables to which to apply the value
586 labels (@pxref{Dictionary Index}). There are @code{var_count}
589 String variables wider than 8 bytes may not be specified in this list.
592 @node Document Record
593 @section Document Record
595 The document record, if present, has the following format:
604 @item int32 rec_type;
605 Record type. Always set to 6.
608 Number of lines of documents present. This should be greater than
609 zero, but the system file writer that identifies itself as
610 @url{https://github.com/WizardMac/ReadStat} writes document records
611 with zero @code{n_lines}.
613 @item char lines[][80];
614 Document lines. The number of elements is defined by @code{n_lines}.
615 Lines shorter than 80 characters are padded on the right with spaces.
618 @node Machine Integer Info Record
619 @section Machine Integer Info Record
621 The integer info record, if present, has the following format:
633 int32 version_revision;
635 int32 floating_point_rep;
636 int32 compression_code;
638 int32 character_code;
642 @item int32 rec_type;
643 Record type. Always set to 7.
646 Record subtype. Always set to 3.
649 Size of each piece of data in the data part, in bytes. Always set to 4.
652 Number of pieces of data in the data part. Always set to 8.
654 @item int32 version_major;
655 PSPP major version number. In version @var{x}.@var{y}.@var{z}, this
658 @item int32 version_minor;
659 PSPP minor version number. In version @var{x}.@var{y}.@var{z}, this
662 @item int32 version_revision;
663 PSPP version revision number. In version @var{x}.@var{y}.@var{z},
666 @item int32 machine_code;
667 Machine code. PSPP always set this field to value to -1, but other
670 @item int32 floating_point_rep;
671 Floating point representation code. For IEEE 754 systems this is 1.
672 IBM 370 sets this to 2, and DEC VAX E to 3.
674 @item int32 compression_code;
675 Compression code. Always set to 1, regardless of whether or how the
678 @item int32 endianness;
679 Machine endianness. 1 indicates big-endian, 2 indicates little-endian.
681 @item int32 character_code;
682 @anchor{character-code} Character code. The following values have
683 been actually observed in system files:
693 The @code{windows-1250} code page for Central European and Eastern
697 The @code{windows-1252} code page for Western European languages.
706 The following additional values are known to be defined:
716 Other Windows code page numbers are known to be generally valid.
718 Old versions of SPSS for Unix and Windows always wrote value 2 in this
719 field, regardless of the encoding in use. Newer versions also write
720 the character encoding as a string (see @ref{Character Encoding
724 @node Machine Floating-Point Info Record
725 @section Machine Floating-Point Info Record
727 The floating-point info record, if present, has the following format:
743 @item int32 rec_type;
744 Record type. Always set to 7.
747 Record subtype. Always set to 4.
750 Size of each piece of data in the data part, in bytes. Always set to 8.
753 Number of pieces of data in the data part. Always set to 3.
756 @itemx flt64 highest;
758 The system missing value, the value used for HIGHEST in missing
759 values, and the value used for LOWEST in missing values, respectively.
760 @xref{System File Format}, for more information.
762 The SPSSWriter library in PHP, which identifies itself as @code{FOM
763 SPSS 1.0.0} in the file header record @code{prod_name} field, writes
764 unexpected values to these fields, but it uses the same values
765 consistently throughout the rest of the file.
768 @node Multiple Response Sets Records
769 @section Multiple Response Sets Records
771 The system file format has two different types of records that
772 represent multiple response sets (@pxref{MRSETS,,,pspp, PSPP Users
773 Guide}). The first type of record describes multiple response sets
774 that can be understood by SPSS before version 14. The second type of
775 record, with a closely related format, is used for multiple dichotomy
776 sets that use the CATEGORYLABELS=COUNTEDVALUES feature added in
786 /* @r{Exactly @code{count} bytes of data.} */
791 @item int32 rec_type;
792 Record type. Always set to 7.
795 Record subtype. Set to 7 for records that describe multiple response
796 sets understood by SPSS before version 14, or to 19 for records that
797 describe dichotomy sets that use the CATEGORYLABELS=COUNTEDVALUES
798 feature added in version 14.
801 The size of each element in the @code{mrsets} member. Always set to 1.
804 The total number of bytes in @code{mrsets}.
807 Zero or more line feeds (byte 0x0a), followed by a series of multiple
808 response sets, each of which consists of the following:
812 The set's name (an identifier that begins with @samp{$}), in mixed
813 upper and lower case.
816 An equals sign (@samp{=}).
819 @samp{C} for a multiple category set, @samp{D} for a multiple
820 dichotomy set with CATEGORYLABELS=VARLABELS, or @samp{E} for a
821 multiple dichotomy set with CATEGORYLABELS=COUNTEDVALUES.
824 For a multiple dichotomy set with CATEGORYLABELS=COUNTEDVALUES, a
825 space, followed by a number expressed as decimal digits, followed by a
826 space. If LABELSOURCE=VARLABEL was specified on MRSETS, then the
827 number is 11; otherwise it is 1.@footnote{This part of the format may
828 not be fully understood, because only a single example of each
829 possibility has been examined.}
832 For either kind of multiple dichotomy set, the counted value, as a
833 positive integer count specified as decimal digits, followed by a
834 space, followed by as many string bytes as specified in the count. If
835 the set contains numeric variables, the string consists of the counted
836 integer value expressed as decimal digits. If the set contains string
837 variables, the string contains the counted string value. Either way,
838 the string may be padded on the right with spaces (older versions of
839 SPSS seem to always pad to a width of 8 bytes; newer versions don't).
845 The multiple response set's label, using the same format as for the
846 counted value for multiple dichotomy sets. A string of length 0 means
847 that the set does not have a label. A string of length 0 is also
848 written if LABELSOURCE=VARLABEL was specified.
854 The short names of the variables in the set, converted to lowercase,
855 each separated from the previous by a single space.
857 Even though a multiple response set must have at least two variables,
858 some system files contain multiple response sets with no variables or
859 one variable. The source and meaning of these multiple response sets is
860 unknown. (Perhaps they arise from creating a multiple response set
861 then deleting all the variables that it contains?)
864 One line feed (byte 0x0a). Sometimes multiple, even hundreds, of line
869 Example: Given appropriate variable definitions, consider the
870 following MRSETS command:
873 MRSETS /MCGROUP NAME=$a LABEL='my mcgroup' VARIABLES=a b c
874 /MDGROUP NAME=$b VARIABLES=g e f d VALUE=55
875 /MDGROUP NAME=$c LABEL='mdgroup #2' VARIABLES=h i j VALUE='Yes'
876 /MDGROUP NAME=$d LABEL='third mdgroup' CATEGORYLABELS=COUNTEDVALUES
877 VARIABLES=k l m VALUE=34
878 /MDGROUP NAME=$e CATEGORYLABELS=COUNTEDVALUES LABELSOURCE=VARLABEL
879 VARIABLES=n o p VALUE='choice'.
882 The above would generate the following multiple response set record of
886 $a=C 10 my mcgroup a b c
888 $c=D3 Yes 10 mdgroup #2 h i j
891 It would also generate the following multiple response set record with
895 $d=E 1 2 34 13 third mdgroup k l m
896 $e=E 11 6 choice 0 n o p
899 @node Extra Product Info Record
900 @section Extra Product Info Record
902 This optional record appears to contain a text string that describes
903 the program that wrote the file and the source of the data. (This is
904 redundant with the file label and product info found in the file
914 /* @r{Exactly @code{count} bytes of data.} */
919 @item int32 rec_type;
920 Record type. Always set to 7.
923 Record subtype. Always set to 10.
926 The size of each element in the @code{info} member. Always set to 1.
929 The total number of bytes in @code{info}.
932 A text string. A product that identifies itself as @code{VOXCO
933 INTERVIEWER 4.3} uses CR-only line ends in this field, rather than the
934 more usual LF-only or CR LF line ends.
937 @node Variable Display Parameter Record
938 @section Variable Display Parameter Record
940 The variable display parameter record, if present, has the following
950 /* @r{Repeated @code{count} times}. */
952 int32 width; /* @r{Not always present.} */
957 @item int32 rec_type;
958 Record type. Always set to 7.
961 Record subtype. Always set to 11.
964 The size of @code{int32}. Always set to 4.
967 The number of sets of variable display parameters (ordinarily the
968 number of variables in the dictionary), times 2 or 3.
971 The remaining members are repeated @code{count} times, in the same
972 order as the variable records. No element corresponds to variable
973 records that continue long string variables. The meanings of these
974 members are as follows:
978 The measurement type of the variable:
988 SPSS sometimes writes a @code{measure} of 0. PSPP interprets this as
992 The width of the display column for the variable in characters.
994 This field is present if @var{count} is 3 times the number of
995 variables in the dictionary. It is omitted if @var{count} is 2 times
996 the number of variables.
998 @item int32 alignment;
999 The alignment of the variable for display purposes:
1011 @node Long Variable Names Record
1012 @section Long Variable Names Record
1014 If present, the long variable names record has the following format:
1023 /* @r{Exactly @code{count} bytes of data.} */
1024 char var_name_pairs[];
1028 @item int32 rec_type;
1029 Record type. Always set to 7.
1031 @item int32 subtype;
1032 Record subtype. Always set to 13.
1035 The size of each element in the @code{var_name_pairs} member. Always set to 1.
1038 The total number of bytes in @code{var_name_pairs}.
1040 @item char var_name_pairs[];
1041 A list of @var{key}--@var{value} tuples, where @var{key} is the name
1042 of a variable, and @var{value} is its long variable name.
1043 The @var{key} field is at most 8 bytes long and must match the
1044 name of a variable which appears in the variable record (@pxref{Variable
1046 The @var{value} field is at most 64 bytes long.
1047 The @var{key} and @var{value} fields are separated by a @samp{=} byte.
1048 Each tuple is separated by a byte whose value is 09. There is no
1049 trailing separator following the last tuple.
1050 The total length is @code{count} bytes.
1053 @node Very Long String Record
1054 @section Very Long String Record
1056 Old versions of SPSS limited string variables to a width of 255 bytes.
1057 For backward compatibility with these older versions, the system file
1058 format represents a string longer than 255 bytes, called a @dfn{very
1059 long string}, as a collection of strings no longer than 255 bytes
1060 each. The strings concatenated to make a very long string are called
1061 its @dfn{segments}; for consistency, variables other than very long
1062 strings are considered to have a single segment.
1064 A very long string with a width of @var{w} has @var{n} =
1065 (@var{w} + 251) / 252 segments, that is, one segment for every
1066 252 bytes of width, rounding up. It would be logical, then, for each
1067 of the segments except the last to have a width of 252 and the last
1068 segment to have the remainder, but this is not the case. In fact,
1069 each segment except the last has a width of 255 bytes. The last
1070 segment has width @var{w} - (@var{n} - 1) * 252; some versions
1071 of SPSS make it slightly wider, but not wide enough to make the last
1072 segment require another 8 bytes of data.
1074 Data is packed tightly into segments of a very long string, 255 bytes
1075 per segment. Because 255 bytes of segment data are allocated for
1076 every 252 bytes of the very long string's width (approximately), some
1077 unused space is left over at the end of the allocated segments. Data
1078 in unused space is ignored.
1080 Example: Consider a very long string of width 20,000. Such a very
1081 long string has 20,000 / 252 = 80 (rounding up) segments. The first
1082 79 segments have width 255; the last segment has width 20,000 - 79 *
1083 252 = 92 or slightly wider (up to 96 bytes, the next multiple of 8).
1084 The very long string's data is actually stored in the 19,890 bytes in
1085 the first 78 segments, plus the first 110 bytes of the 79th segment
1086 (19,890 + 110 = 20,000). The remaining 145 bytes of the 79th segment
1087 and all 92 bytes of the 80th segment are unused.
1089 The very long string record explains how to stitch together segments
1090 to obtain very long string data. For each of the very long string
1091 variables in the dictionary, it specifies the name of its first
1092 segment's variable and the very long string variable's actual width.
1093 The remaining segments immediately follow the named variable in the
1094 system file's dictionary.
1096 The very long string record, which is present only if the system file
1097 contains very long string variables, has the following format:
1106 /* @r{Exactly @code{count} bytes of data.} */
1107 char string_lengths[];
1111 @item int32 rec_type;
1112 Record type. Always set to 7.
1114 @item int32 subtype;
1115 Record subtype. Always set to 14.
1118 The size of each element in the @code{string_lengths} member. Always set to 1.
1121 The total number of bytes in @code{string_lengths}.
1123 @item char string_lengths[];
1124 A list of @var{key}--@var{value} tuples, where @var{key} is the name
1125 of a variable, and @var{value} is its length.
1126 The @var{key} field is at most 8 bytes long and must match the
1127 name of a variable which appears in the variable record (@pxref{Variable
1129 The @var{value} field is exactly 5 bytes long. It is a zero-padded,
1130 ASCII-encoded string that is the length of the variable.
1131 The @var{key} and @var{value} fields are separated by a @samp{=} byte.
1132 Tuples are delimited by a two-byte sequence @{00, 09@}.
1133 After the last tuple, there may be a single byte 00, or @{00, 09@}.
1134 The total length is @code{count} bytes.
1137 @node Character Encoding Record
1138 @section Character Encoding Record
1140 This record, if present, indicates the character encoding for string data,
1141 long variable names, variable labels, value labels and other strings in the
1151 /* @r{Exactly @code{count} bytes of data.} */
1156 @item int32 rec_type;
1157 Record type. Always set to 7.
1159 @item int32 subtype;
1160 Record subtype. Always set to 20.
1163 The size of each element in the @code{encoding} member. Always set to 1.
1166 The total number of bytes in @code{encoding}.
1168 @item char encoding[];
1169 The name of the character encoding. Normally this will be an official
1170 IANA character set name or alias.
1171 See @url{http://www.iana.org/assignments/character-sets}.
1172 Character set names are not case-sensitive, but SPSS appears to write
1173 them in all-uppercase.
1176 This record is not present in files generated by older software. See
1177 also the @code{character_code} field in the machine integer info
1178 record (@pxref{character-code}).
1180 When the character encoding record and the machine integer info record
1181 are both present, all system files observed in practice indicate the
1182 same character encoding, e.g.@: 1252 as @code{character_code} and
1183 @code{windows-1252} as @code{encoding}, 65001 and @code{UTF-8}, etc.
1185 If, for testing purposes, a file is crafted with different
1186 @code{character_code} and @code{encoding}, it seems that
1187 @code{character_code} controls the encoding for all strings in the
1188 system file before the dictionary termination record, including
1189 strings in data (e.g.@: string missing values), and @code{encoding}
1190 controls the encoding for strings following the dictionary termination
1193 @node Long String Value Labels Record
1194 @section Long String Value Labels Record
1196 This record, if present, specifies value labels for long string
1206 /* @r{Repeated up to exactly @code{count} bytes.} */
1211 long_string_label labels[];
1215 @item int32 rec_type;
1216 Record type. Always set to 7.
1218 @item int32 subtype;
1219 Record subtype. Always set to 21.
1225 The number of bytes following the header until the next header.
1227 @item int32 var_name_len;
1228 @itemx char var_name[];
1229 The number of bytes in the name of the variable that has long string
1230 value labels, plus the variable name itself, which consists of exactly
1231 @code{var_name_len} bytes. The variable name is not padded to any
1232 particular boundary, nor is it null-terminated.
1234 @item int32 var_width;
1235 The width of the variable, in bytes, which will be between 9 and
1238 @item int32 n_labels;
1239 @itemx long_string_label labels[];
1240 The long string labels themselves. The @code{labels} array contains
1241 exactly @code{n_labels} elements, each of which has the following
1252 @item int32 value_len;
1253 @itemx char value[];
1254 The string value being labeled. @code{value_len} is the number of
1255 bytes in @code{value}; it is equal to @code{var_width}. The
1256 @code{value} array is not padded or null-terminated.
1258 @item int32 label_len;
1259 @itemx char label[];
1260 The label for the string value. @code{label_len}, which must be
1261 between 0 and 120, is the number of bytes in @code{label}. The
1262 @code{label} array is not padded or null-terminated.
1266 @node Long String Missing Values Record
1267 @section Long String Missing Values Record
1269 This record, if present, specifies missing values for long string
1279 /* @r{Repeated up to exactly @code{count} bytes.} */
1282 char n_missing_values;
1283 long_string_missing_value values[];
1287 @item int32 rec_type;
1288 Record type. Always set to 7.
1290 @item int32 subtype;
1291 Record subtype. Always set to 22.
1297 The number of bytes following the header until the next header.
1299 @item int32 var_name_len;
1300 @itemx char var_name[];
1301 The number of bytes in the name of the long string variable that has
1302 missing values, plus the variable name itself, which consists of
1303 exactly @code{var_name_len} bytes. The variable name is not padded to
1304 any particular boundary, nor is it null-terminated.
1306 @item char n_missing_values;
1307 The number of missing values, either 1, 2, or 3. (This is, unusually,
1308 a single byte instead of a 32-bit number.)
1310 @item long_string_missing_value values[];
1311 The missing values themselves. This array contains exactly
1312 @code{n_missing_values} elements, each of which has the following
1321 @item int32 value_len;
1322 The length of the missing value string, in bytes. This value should
1323 be 8, because long string variables are at least 8 bytes wide (by
1324 definition), only the first 8 bytes of a long string variable's
1325 missing values are allowed to be non-spaces, and any spaces within the
1326 first 8 bytes are included in the missing value here.
1329 The missing value string, exactly @code{value_len} bytes, without
1330 any padding or null terminator.
1334 @node Data File and Variable Attributes Records
1335 @section Data File and Variable Attributes Records
1337 The data file and variable attributes records represent custom
1338 attributes for the system file or for individual variables in the
1339 system file, as defined on the DATAFILE ATTRIBUTE (@pxref{DATAFILE
1340 ATTRIBUTE,,,pspp, PSPP Users Guide}) and VARIABLE ATTRIBUTE commands
1341 (@pxref{VARIABLE ATTRIBUTE,,,pspp, PSPP Users Guide}), respectively.
1350 /* @r{Exactly @code{count} bytes of data.} */
1355 @item int32 rec_type;
1356 Record type. Always set to 7.
1358 @item int32 subtype;
1359 Record subtype. Always set to 17 for a data file attribute record or
1360 to 18 for a variable attributes record.
1363 The size of each element in the @code{attributes} member. Always set to 1.
1366 The total number of bytes in @code{attributes}.
1368 @item char attributes[];
1369 The attributes, in a text-based format.
1371 In record subtype 17, this field contains a single attribute set. An
1372 attribute set is a sequence of one or more attributes concatenated
1373 together. Each attribute consists of a name, which has the same
1374 syntax as a variable name, followed by, inside parentheses, a sequence
1375 of one or more values. Each value consists of a string enclosed in
1376 single quotes (@code{'}) followed by a line feed (byte 0x0a). A value
1377 may contain single quote characters, which are not themselves escaped
1378 or quoted or required to be present in pairs. There is no apparent
1379 way to embed a line feed in a value. There is no distinction between
1380 an attribute with a single value and an attribute array with one
1383 In record subtype 18, this field contains a sequence of one or more
1384 variable attribute sets. If more than one variable attribute set is
1385 present, each one after the first is delimited from the previous by
1386 @code{/}. Each variable attribute set consists of a long
1388 followed by @code{:}, followed by an attribute set with the same
1389 syntax as on record subtype 17.
1391 System files written by @code{Stata 14.1/-savespss- 1.77 by
1392 S.Radyakin} may include multiple records with subtype 18, one per
1393 variable that has variable attributes.
1395 The total length is @code{count} bytes.
1400 A system file produced with the following VARIABLE ATTRIBUTE commands
1404 VARIABLE ATTRIBUTE VARIABLES=dummy ATTRIBUTE=fred[1]('23') fred[2]('34').
1405 VARIABLE ATTRIBUTE VARIABLES=dummy ATTRIBUTE=bert('123').
1409 will contain a variable attribute record with the following contents:
1412 0000 07 00 00 00 12 00 00 00 01 00 00 00 22 00 00 00 |............"...|
1413 0010 64 75 6d 6d 79 3a 66 72 65 64 28 27 32 33 27 0a |dummy:fred('23'.|
1414 0020 27 33 34 27 0a 29 62 65 72 74 28 27 31 32 33 27 |'34'.)bert('123'|
1422 @node Variable Roles
1423 @subsection Variable Roles
1425 A variable's role is represented as an attribute named @code{$@@Role}.
1426 This attribute has a single element whose values and their meanings
1431 Input. This, the default, is the most common role.
1444 @node Extended Number of Cases Record
1445 @section Extended Number of Cases Record
1447 The file header record expresses the number of cases in the system
1448 file as an int32 (@pxref{File Header Record}). This record allows the
1449 number of cases in the system file to be expressed as a 64-bit number.
1461 @item int32 rec_type;
1462 Record type. Always set to 7.
1464 @item int32 subtype;
1465 Record subtype. Always set to 16.
1468 Size of each element. Always set to 8.
1471 Number of pieces of data in the data part. Alway set to 2.
1473 @item int64 unknown;
1474 Meaning unknown. Always set to 1.
1476 @item int64 ncases64;
1477 Number of cases in the file as a 64-bit integer. Presumably this
1478 could be -1 to indicate that the number of cases is unknown, for the
1479 same reason as @code{ncases} in the file header record, but this has
1480 not been observed in the wild.
1483 @node Other Informational Records
1484 @section Other Informational Records
1486 This chapter documents many specific types of extension records are
1487 documented here, but others are known to exist. PSPP ignores unknown
1488 extension records when reading system files.
1490 The following extension record subtypes have also been observed, with
1491 the following believed meanings:
1495 A set of grouped variables (according to Aapi H@"am@"al@"ainen).
1498 Date info, probably related to USE (according to Aapi H@"am@"al@"ainen).
1501 A UUID in the format described in RFC 4122. Only two examples
1502 observed, both written by SPSS 13, and in each case the UUID contained
1503 both upper and lower case.
1506 XML that describes how data in the file should be displayed on-screen.
1509 @node Dictionary Termination Record
1510 @section Dictionary Termination Record
1512 The dictionary termination record separates all other records from the
1521 @item int32 rec_type;
1522 Record type. Always set to 999.
1525 Ignored padding. Should be set to 0.
1529 @section Data Record
1531 The data record must follow all other records in the system file.
1532 Every system file must have a data record that specifies data for at
1533 least one case. The format of the data record varies depending on the
1534 value of @code{compression} in the file header record:
1537 @item 0: no compression
1538 Data is arranged as a series of 8-byte elements.
1539 Each element corresponds to
1540 the variable declared in the respective variable record (@pxref{Variable
1541 Record}). Numeric values are given in @code{flt64} format; string
1542 values are literal characters string, padded on the right when
1543 necessary to fill out 8-byte units.
1545 @item 1: bytecode compression
1547 of the data record is divided into a series of 1-byte command
1548 codes. These codes have meanings as described below:
1552 Ignored. If the program writing the system file accumulates compressed
1553 data in blocks of fixed length, 0 bytes can be used to pad out extra
1554 bytes remaining at the end of a fixed-size block.
1558 value @var{code} - @var{bias}, where
1559 @var{code} is the value of the compression code and @var{bias} is the
1560 variable @code{bias} from the file header. For example,
1561 code 105 with bias 100.0 (the normal value) indicates a numeric variable
1563 One file has been seen written by SPSS 14 that contained such a code
1564 in a @emph{string} field with the value 0 (after the bias is
1565 subtracted) as a way of encoding null bytes.
1568 End of file. This code may or may not appear at the end of the data
1569 stream. PSPP always outputs this code but its use is not required.
1572 A numeric or string value that is not
1573 compressible. The value is stored in the 8 bytes following the
1574 current block of command bytes. If this value appears twice in a block
1575 of command bytes, then it indicates the second group of 8 bytes following the
1576 command bytes, and so on.
1579 An 8-byte string value that is all spaces.
1582 The system-missing value.
1585 The end of the 8-byte group of bytecodes is followed by any 8-byte
1586 blocks of non-compressible values indicated by code 253. After that
1587 follows another 8-byte group of bytecodes, then those bytecodes'
1588 non-compressible values. The pattern repeats to the end of the file
1589 or a code with value 252.
1591 @item 2: ZLIB compression
1592 The data record consists of the following, in order:
1596 ZLIB data header, 24 bytes long.
1599 One or more variable-length blocks of ZLIB compressed data.
1602 ZLIB data trailer, with a 24-byte fixed header plus an additional 24
1603 bytes for each preceding ZLIB compressed data block.
1606 The ZLIB data header has the following format:
1615 @item int64 zheader_ofs;
1616 The offset, in bytes, of the beginning of this structure within the
1619 @item int64 ztrailer_ofs;
1620 The offset, in bytes, of the first byte of the ZLIB data trailer.
1622 @item int64 ztrailer_len;
1623 The number of bytes in the ZLIB data trailer. This and the previous
1624 field sum to the size of the system file in bytes.
1627 The data header is followed by @code{(ztrailer_ofs - 24) / 24} ZLIB
1628 compressed data blocks. Each ZLIB compressed data block begins with a
1629 ZLIB header as specified in RFC@tie{}1950, e.g.@: hex bytes @code{78
1630 01} (the only header yet observed in practice). Each block
1631 decompresses to a fixed number of bytes (in practice only
1632 @code{0x3ff000}-byte blocks have been observed), except that the last
1633 block of data may be shorter. The last ZLIB compressed data block
1634 gends just before offset @code{ztrailer_ofs}.
1636 The result of ZLIB decompression is bytecode compressed data as
1637 described above for compression format 1.
1639 The ZLIB data trailer begins with the following 24-byte fixed header:
1649 @item int64 int_bias;
1650 The compression bias as a negative integer, e.g.@: if @code{bias} in
1651 the file header record is 100.0, then @code{int_bias} is @minus{}100
1652 (this is the only value yet observed in practice).
1655 Always observed to be zero.
1657 @item int32 block_size;
1658 The number of bytes in each ZLIB compressed data block, except
1659 possibly the last, following decompression. Only @code{0x3ff000} has
1660 been observed so far.
1662 @item int32 n_blocks;
1663 The number of ZLIB compressed data blocks, always exactly
1664 @code{(ztrailer_ofs - 24) / 24}.
1667 The fixed header is followed by @code{n_blocks} 24-byte ZLIB data
1668 block descriptors, each of which describes the compressed data block
1669 corresponding to its offset. Each block descriptor has the following
1673 int64 uncompressed_ofs;
1674 int64 compressed_ofs;
1675 int32 uncompressed_size;
1676 int32 compressed_size;
1680 @item int64 uncompressed_ofs;
1681 The offset, in bytes, that this block of data would have in a similar
1682 system file that uses compression format 1. This is
1683 @code{zheader_ofs} in the first block descriptor, and in each
1684 succeeding block descriptor it is the sum of the previous desciptor's
1685 @code{uncompressed_ofs} and @code{uncompressed_size}.
1687 @item int64 compressed_ofs;
1688 The offset, in bytes, of the actual beginning of this compressed data
1689 block. This is @code{zheader_ofs + 24} in the first block descriptor,
1690 and in each succeeding block descriptor it is the sum of the previous
1691 descriptor's @code{compressed_ofs} and @code{compressed_size}. The
1692 final block descriptor's @code{compressed_ofs} and
1693 @code{compressed_size} sum to @code{ztrailer_ofs}.
1695 @item int32 uncompressed_size;
1696 The number of bytes in this data block, after decompression. This is
1697 @code{block_size} in every data block except the last, which may be
1700 @item int32 compressed_size;
1701 The number of bytes in this data block, as stored compressed in this
1706 @setfilename ignored