JPH03245233A - Label retrieving system - Google Patents

Abstract

PURPOSE:To improve the check efficiency by producing a 2nd hash table when the labels (synonym) having the same hash values obtained by the label are equal to a prescribed number. CONSTITUTION:The label Namej taken out via a label reading means 1 is registered. The hash value mj is produced from a label as mj = Hm(Namej), where Hm shows a hash function. If the number of synonyms contained in a 1st hash table 71 corresponding to the value mj is equal to the number of arrays of tables, a 2nd hash table production means 4 registers again the table defining the value mj as a synonym with use of Hn. A table register means 5 retrieves the tables 7 via the mj, registers the Namej in a prescribed place, and adds '1' to the synonym number of the table 71. A registration/retrieval deciding means 2 starts a synonym number retrieving means 3 at registration of labels and then starts a table retrieving means 6 at retrieval of labels respectively based on the syntax rule of a programming language. Then the label retrieving efficiency is improved owing to the reduction of the number of synonyms even though many labels having the same hash values due to the labels may emerge in a large number.

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to a name search method in programming language translation processing.

[Conventional technology]

Conventionally, this type of name search method uses one type of hash table to perform registration in the hash table and search thereof.

[Problem to be solved by the invention]

In the conventional name search method described above, when names (synonyms) with the same hash value appear, they are registered in the same synonym chain from the hash chain on the hash table, no matter how many times they appear. Registration of marks/
If there are a large number of synonyms in the search process at the time of reference, a large number of names that do not need to be searched will be referenced, resulting in a decrease in search effectiveness.

[Means to solve the problem]

The name search method of the present invention is a name search method in a programming language translation process, which includes a name reading means for extracting a name from a source program, and a first hash table when registering a name. synonym number checking means for checking whether the number of synonyms (names with the same hash value) is equal to a determined number; and creating a second hash table when the number of synonyms is equal to a determined number. Second
a hash table creating means; a table registration means for registering the name name obtained by the name name reading means in the first and second hash tables; and a hash table in which the name is registered when referring to the name mark. and a table search means for searching from.

〔Example〕

Next, the present invention will be explained with reference to the drawings.

FIG. 1 is a block diagram of an embodiment of the present invention. Referring to FIG. 1, this embodiment includes a name reading means 1 for extracting a name from a source program 8, and a name reading means 1 to check whether the number of synonyms in the first hash table 71 is equal to a predetermined number when registering a name. A synonym number checking means 3 to check, a second hash table creation means 4 for creating a second hash table 72 when the number of synonyms is equal to a predetermined number, and a table registration means 5 for registering in the name tag hash table 7. When referring to a name tag, a table search means 6 searches from the hash table 7 in which the name name is registered, a hash table 7 registers the name name, and a name reading means 1 selects the name name to be retrieved. It consists of a registration/search determination means 2 which determines whether it is a registration or a reference and activates a synonym check means 3 or a table search means 6. Furthermore, the hash table 7 includes the above-mentioned first hash table 71 and second hash table 72.
have.

FIG. 2 is a state diagram of registration in the hash table 7 of FIG. 1, and registration of a name mark will be described with reference to FIG. Let us consider the case where the name name Namej retrieved by the name name reading means 1 is to be registered. The hash value mJ is generated by the name tag.

m, = Hm (Namej) (H, is a hash function) In this case, if the number of synonyms of the first hash table corresponding to "m" becomes equal to a fixed number (for example, the number of arrays of the first hash table) , the second hash table creation means 4 generates a hash function H, , (H, is Hl, l
The second hash table is re-registered for the names (excluding Namej) that have f'rlJ as a synonym. The table registration means 5 searches the hash table 7 using the hash value mJ and registers the name name Namej in a predetermined location (currently, the number of synonyms in the first hash table corresponding to m is determined). Since the numbers are similar, the number of synonyms in the first hash table corresponding to m (name name Namej is chained from the second hash table) is counted up by one.

Hash values mI, . . . ml were generated and registered in the hash table 7 up to Name 1, . . . Namei. When a name tag with a hash value equal to that gap appears, the name tag is registered using a synonym chain from the first hash tail pull.

FIG. 2(a) shows the state of the hash table 7 before the name name Namej is registered, that is, before the second hash table corresponding to m,1 is created, and FIG. 2(b) shows the state of the hash table 7 before the name name Namej is registered.
The state of the hash table 7 is shown after a m e j is registered, that is, after the second hash table corresponding to m is created.

Next, we will discuss when referring to name tags. When referring to the name name Namej, the table search means 6 uses the hash interval number Hf
The first hash table is searched using the hash value mJ generated from fi. (In this case, since the number of synonyms in the first hash table corresponding to mj exceeds the predetermined number, the hash value m
Search the second hash table using J. )In addition,
The registration/search determination means 2 follows the syntax rules of the programming language, and activates the synonym number search means 3 when registering a name, and the table search means 6 when searching for a name.

[Effects of the Invention] As explained above, the present invention creates a second hash table when the halash value of a name tag is equal to the number of identical name tags. Even if many items with the same value appear, synonyms are reduced and search efficiency is improved.

BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a block diagram of an embodiment of the present invention, and FIG. 2 is a state diagram of name registration.

1...Name tag reading means, 2...Registration/
Search determination means, 3...Synonym number inspection means, 4
...Second hash table creation means, 5...
... table registration means, 6 ... table search means, 7 ... hash table, 71 ...
...First hash table, 72...Second hash table, 8...Source program.

Claims (1)

[Claims] In the search method for name names in the translation process of programming languages, there is a name reading means for extracting name names from a source program, and synonyms (name names with the same hash value) in the first hash table when registering name names. ) synonym number checking means for checking whether the number is equal to a predetermined number; second hash table creation means for creating a second hash table when the number of synonyms is equal to the predetermined number; and the name tag reading means. A name characterized by comprising table registration means for registering the obtained name name in the first and second hash tables, and table search means for searching from the hash table in which the name is registered when referring to the name name. Mark search method.