International
Tables for
Crystallography
Volume G
Definition and exchange of crystallographic data
Edited by S. R. Hall and B. McMahon

International Tables for Crystallography (2006). Vol. G, ch. 2.1, pp. 14-16

Section 2.1.3. The syntax of the STAR File

S. R. Halla* and N. Spadaccinib

aSchool of Biomedical and Chemical Sciences, University of Western Australia, Crawley, Perth, WA 6009, Australia, and bSchool of Computer Science and Software Engineering, University of Western Australia, 35 Stirling Highway, Crawley, Perth, WA 6009, Australia
Correspondence e-mail:  syd@crystal.uwa.edu.au

2.1.3. The syntax of the STAR File

| top | pdf |

The syntax of the STAR File (Hall, 1991[link]; Hall & Spadaccini, 1994[link]) has been used to develop a number of discipline-specific exchange and archival approaches, including the Crystallographic Information File (CIF) (Hall et al., 1991[link]), the Molecular Information File (MIF) (Allen et al., 1995[link]), the dictionary definition language (DDL1) (Hall & Cook, 1995[link]), the macromolecular dictionary definition language (DDL2) (Westbrook & Hall, 1995[link]) and the STAR dictionary definition language (StarDDL) (Spadaccini et al., 2000[link]). The details of the CIF, MIF, DDL1 and DDL2 approaches are given in Chapters 2.2[link] , 2.4[link] , 2.5[link] and 2.6[link] , respectively.

A STAR File is a sequential file containing lines of standard ASCII characters. A file may be divided into any number of discrete sets of unique data items. Sets may be in the form of data blocks, global blocks or save frames. The syntax rules for these sets are given below in descriptive form. A more rigorous description of the STAR File syntax is given in Appendix 2.1.1[link] in extended Backus–Naur form (McLennon, 1983[link]).

The STAR File is a free-form language in which spaces (ASCII 32), vertical tabs (ASCII 11) and horizontal tabs (ASCII 9) are collectively referred to as <blank>, and newlines (ASCII 10), form feeds (ASCII 12) and carriage returns (ASCII 13) are collectively referred to as <terminate>. White spaces <wspace>, used to separate lexical tokens within the file, are all characters in the joined set of <blank> and <terminate>.

2.1.3.1. Text string

| top | pdf |

A text string is defined as any of the following.

  • (a) A sequence of non-white-space characters on a single line excluding a leading underscore <_> (ASCII 95).

    Examples: [Scheme scheme1]

  • (b) A sequence of characters on a single line containing the leading digraph <wspace><'> and the trailing digraph <'><wspace>. <'> is a single-quote character (ASCII 39) and <wspace> is white space.

    Examples: [Scheme scheme2]

    Note that the use of the <'> character in the text string that is bounded by a <'> character is not precluded unless it is immediately followed by <wspace>. The leading and trailing digraphs serve to delimit the string and do not form part of the data. In the above example the value associated with the text field 'light blue' is light blue.

  • (c) A sequence of characters on a single line containing the leading digraph <wspace><"> and the trailing digraph <"><wspace>. <"> is a double-quote (ASCII 34) character and <wspace> is white space.

    Examples: [Scheme scheme3]

    The use of the <"> character in the text string that is bounded by a <"> character is not precluded unless it is immediately followed by <wspace>. The leading and trailing digraphs serve to delimit the string and do not form part of the data.

    The text strings of type (a), (b) and (c) cannot contain line-breaking characters, and therefore cannot span multiple lines. There can be more than one text string per line if each value is preceded by a data name, or the values are part of a looped list (see Section 2.1.3.5[link]).

  • (d) A sequence of lines starting with <terminate><;> and finishing with <terminate><;>, where <;> is the semicolon character (ASCII 59).

    Example: [Scheme scheme4]

    The requirement that the <;> character be the first character on the line does not preclude the presence of other characters on the same line, in as much as it does not violate the STAR File structure.

    The leading and trailing digraphs delimit the text field and do not form part of the data. The character sequence between the digraphs, including any line-breaking characters, constitutes the value of the text field. The value associated with the above example is <blank>School<blank>of<blank>CSSE<terminate><blank><blank>UWA (note in particular that the <terminate> character preceding the final ; delimiter is not part of the value).

2.1.3.2. Data name

| top | pdf |

A data name (or tag) is the identifier of a data value (see Section 2.1.3.3[link]) and is a sequence of non-white-space characters starting with an underscore character <_> (ASCII 95).

Example: [Scheme scheme5]

2.1.3.3. Data value

| top | pdf |

A data value is a text string preceded by its identifying data name. Privileged keywords, such as described in Sections 2.1.3.5[link] to 2.1.3.8[link], are excluded from this definition.

2.1.3.4. Data item

| top | pdf |

A data item is a data value and its associated data name. Each data item stored in a STAR File is specified with this combination.

2.1.3.5. Data loop list

| top | pdf |

A looped list consists of the keyword loop_ followed by

  • (a) a sequence of data names (possibly with nested loop_ constructs); and

  • (b) a sequence of loop packets, each containing data values which are identified in the same order as the data names in (a).

A looped list specifies a table of data in which the data names represent the `header descriptors' for columns of data and the packets represent the rows in the table. Looped lists may be nested to any level. Each loop level is initialized with the loop_ keyword and is followed by the names of data items in this level. Data values that follow the nested data declarations must be in exact multiples of the number of data names. Each loop level must be terminated with a stop_, except the outermost (level 1) which is terminated by either a new data item or the privileged strings indicating a save frame (Section 2.1.3.6[link]), a data block (Section 2.1.3.7[link]), a global block (Section 2.1.3.8[link]) or an end of file.

An example of a simple one-level loop structure is: [Scheme scheme6]

Nested (multi-level) looped lists contain matching data packets [as per (b[link]) above] and an additional stop_ to terminate each level of data. Here is a simple example of a two-level nested list. [Scheme scheme7]

The matching of data names to value packets is applied at each loop level. Initially the data values are matched to the data names listed in the outermost level loop. This process is iterated to successively inner levels. At the innermost loop level, data matching is maintained until a stop_ is encountered. This returns the matching process to the next outer level. The matching process is recursive until the loop structure is depleted. Here is an example of a three-level loop structure. [Scheme scheme8]

2.1.3.6. Save frame

| top | pdf |

A save frame is a set of unique data items wholly contained within a data block. The frame starts with a save_framecode statement, where the framecode is a unique identifying code within the data block. Each frame is closed with a save_ statement.

Example: [Scheme scheme9]

A save frame has the following attributes:

  • (a) A save frame may contain data items and loop structures but not other save frames [see (f[link])].

  • (b) The scope of the data specified in a save frame is the save frame in which it is specified.

  • (c) Data values in a save frame are distinct from any identical items in the parent data block.

  • (d) A save frame may be referenced within the data block in which it is specified using a data item with a value of $framecode.

    Example: [Scheme scheme10] where `arg', `tyr' and `leu' are frame codes identifying three save frames of data.

  • (e) A frame code must be unique within a data block.

  • (f) A save frame may not contain another save frame, but it may contain references to other save frames in the same data block using frame codes.

2.1.3.7. Data block

| top | pdf |

A data block is a set of data containing any number of unique items and save frames. A data block begins with a data_blockcode statement, where blockcode is a unique identifying name within a file. A data block is closed by another data_blockcode statement, a global_ statement or an end of file.

Example: [Scheme scheme11]

all information relevant to rhinovirus included here

[Scheme scheme12]

all information relevant to influenza virus
included here

A data block has the following attributes:

  • (a) A block code must be unique within the file containing the data block.

  • (b) Data blocks may not be referenced from within a file [in contrast to save frames – see Section 2.1.3.6[link](d)].

  • (c) The scope of data specified in a data block is the data block. The value of a data item is always associated with the data block in which it is specified.

  • (d) Data specifications in a data block are unique, except they may be repeated within a save frame. Data specifications in a save frame are independent of the parent data block specifications.

  • (e) If a data item is not specified in a given data block, the global value is assumed. If a global value is not specified, the value is unknown.

2.1.3.8. Global block

| top | pdf |

A global block is a set of data items which are implied to be present in all data blocks which follow in a file, unless specified explicitly within a data block. A global block starts with a global_ keyword and is closed by a data_blockcode statement or an end of file.

Example: [Scheme scheme13]

information that is default within subsequent
data blocks

[Scheme scheme14]

A global block has the following attributes:

  • (a) The scope of global data is from the point of declaration to the end of file.

  • (b) A global block may contain data items, loop structures and save frames.

  • (c) Multiple global blocks are concatenated to form a single block in which the last item specification has precedence.

  • (d) A data item specified within a data block has precedence over a data item specified in a prior global block.

2.1.3.9. Data sets and scopes

| top | pdf |

A data set is the generic term for a unique set of data. A STAR File may contain three types of data sets: global blocks, data blocks and save frames. The attributes of data sets are as follows.

  • (a) A file may contain any number of data sets.

  • (b) The data names defined within a data set must be unique to that set. That is, all data_blockcode names must be unique within the file, all data names must be unique within a global_ block, all data names and save_framecodes must be unique within a data block, and all data names must be unique within a save frame.

  • (c) The scope of data sets is hierarchical (Fig. 2.1.3.1[link]). Global blocks encompass all following data blocks; data blocks scope all contained save frames.

    [Figure 2.1.3.1]

    Figure 2.1.3.1 | top | pdf |

    Nested scopes of STAR data sets.

  • (d) The scope of a save frame is all data items contained within the frame.

  • (e) The scope of a data block is the boundaries of the data block, i.e. the end of the file or the start of the next data block, including any contained save frames. The same data item may be defined within a save frame and within the parent data block. All specifications of this item will be recognized when accessing the data block.

  • (f) The scope of a global block is the file, from the point of invocation to the end of file or the start of the next global block. It encompasses all contained global data items, data blocks and save frames. Globally specified data are active provided identical items are not specified in subsequent data sets.

2.1.3.10. Privileged constructs

| top | pdf |

The following constructs are privileged.

  • (a) Text strings starting with the character sequences data_, loop_, global_, save_ or stop_ are privileged words (keywords) and may not be used as values in text strings of the type defined in Section 2.1.3.1(a[link]).

  • (b) A sharp character <#> (ASCII 35) is an explicit end-of-line signal provided it is not contained within a text string of the types defined in Section 2.1.3.1(b[link]), (c[link]) or (d[link]). Characters on the same line and following an active sharp character are considered as comment text.

.

2.1.3.11. Using stop_ in looped lists

| top | pdf |

In Section 2.1.3.5[link] we discuss how stop_ is used to terminate a loop of data values and to return the looped list to the next outer nesting level. This same construction applies in the looped list of data names. The following, although not particularly intuitive, is a valid construction. [Scheme scheme15]

This is equivalent to the loop definition given in Section 2.1.3.5[link]. One can use the stop_ in name definitions to inhibit the nesting of loops in the definitions.

References

Allen, F. H., Barnard, J. M., Cook, A. P. F. & Hall, S. R. (1995). The Molecular Information File (MIF): core specifications of a new standard format for chemical data. J. Chem. Inf. Comput. Sci. 35, 412–427.
Hall, S. R. (1991). The STAR File: a new format for electronic data transfer and archiving. J. Chem. Inf. Comput. Sci. 31, 326–333.
Hall, S. R., Allen, F. H. & Brown, I. D. (1991). The Crystallographic Information File (CIF): a new standard archive file for crystallography. Acta Cryst. A47, 655–685.
Hall, S. R. & Cook, A. P. F. (1995). STAR dictionary definition language: initial specification. J. Chem. Inf. Comput. Sci. 35, 819–825.
Hall, S. R. & Spadaccini, N. (1994). The STAR File: detailed specifications. J. Chem. Inf. Comput. Sci. 34, 505–508.
McLennon, B. J. (1983). Principles of programming languages, design, evaluation and implementation. New York: Holt, Rinehart and Winston.
Spadaccini, N., Hall, S. R. & Castleden, I. R. (2000). Relational expressions in STAR File dictionaries. J. Chem. Inf. Comput. Sci. 40, 1289–1301.
Westbrook, J. D. & Hall, S. R. (1995). A Dictionary Description Language for Macromolecular Structure. http://ndbserver.rutgers.edu/mmcif/ddl .








































to end of page
to top of page