CN103365998B - A kind of similar character string search method - Google Patents
A kind of similar character string search method Download PDFInfo
- Publication number
- CN103365998B CN103365998B CN201310294529.4A CN201310294529A CN103365998B CN 103365998 B CN103365998 B CN 103365998B CN 201310294529 A CN201310294529 A CN 201310294529A CN 103365998 B CN103365998 B CN 103365998B
- Authority
- CN
- China
- Prior art keywords
- gram
- character string
- row chain
- index
- search method
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Landscapes
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention discloses a kind of similar character string search method, including: utilize the coincidence information between the table of falling row chain in inverted index, obtain gram identification index；Gram identification index is integrated with the table of falling row chain length ratio, forms aggregative indicator；Obtain Candidate Set according to aggregative indicator by extensive prefix filter, calculate the real editing distance of character string in Candidate Set and obtain retrieval result.Similar character search method of the present invention can improve recall precision, especially when editing distance threshold values increases, it is possible to obtains significant increase recall precision.
Description
Technical field
The invention belongs to database technical field, be specifically related to a kind of method that similar character string data is retrieved.
Background technology
In recent years, database field has carried out numerous studies to string-similarity inquiry.String-similarity inquiry can
To be widely used in data fusion, data cleansing, repeatability detection, in natural language processing and the medium field of bioinformatics.
Such as, may there are some mistakes in the non-homogeneous data-base recording representing identical entity in real world, set up data bins
Need during storehouse these identical entities are carried out data fusion；In protein detection, need the protein sequence that quick obtaining is new
Relevant information, now can use the protein sequence with new protein is most like that new protein sequence is retouched
State；In existing search engine, owing to the input of people usually contains wrong, search engine is now needed to provide some error correction
Function, points out and helps user to identify mistake present in its input.
Existing up-to-date proposition solve string-similarity inquiry method have, VChunk, asymmetric-Search and
Adaptive-Search.Wherein, VChunk utilizes the structure of a kind of Chunk border dictionary based on statistics to enter character string
Row splits the signature generated for retrieval.When asymmetric-Search is by utilizing inquiry to use different signatures with indexing
(during an employing gram, another uses chunk) can reduce occurrence, thus reaches to put forward high performance purpose.
Adaptive-Search is then based primarily upon an observation: found by laboratory observation, the when of sewing filter before use, suitable
Performance boost can be obtained when increasing the gram number for accessing inverted list.Before the basic filter that first two method uses is
Sewing filter, tradition prefix filter can produce substantial amounts of Candidate Set；The third method is intended to by increasing reality for accessing
The gram number of inverted index improves traditional prefix filter, but owing to being based only on the heuristic rule of laboratory observation,
Therefore it is not reaching to best performance.
Summary of the invention
Instant invention overcomes editing distance threshold values increase in prior art causes Candidate Set filter excessive, multiple cannot oneself
The defects such as adaptation selects, the quantizating index of existing query optimization is the most accurate, it is proposed that a kind of similar character string search method.
The present invention proposes a kind of similar character string search method, comprises the following steps:
Step one: utilize the coincidence information between the table of falling row chain in inverted index, obtains gram identification index；
Step 2: described gram identification index integrated with the table of falling row chain length ratio, forms aggregative indicator；
Step 3: obtain Candidate Set by extensive prefix filter according to described aggregative indicator, by calculating described candidate
The real editing distance of character string concentrated obtains retrieval result.
The present invention proposes a kind of similar character string search method, wherein, before step one, is first retrieved needs
String assemble sets up inverted index, and generates the MinHash signature that inverted index is corresponding.
The present invention proposes a kind of similar character string search method, wherein, by being numbered character string and q-gram
Splitting, the gram after splitting sets up inverted index for entry maps to the relation of the character string comprising described gram.
The present invention proposes a kind of similar character string search method, wherein, chooses Hash according to the length of described character string
Function, generates MinHash signature in conjunction with described inverted index.
The present invention proposes a kind of similar character string search method, and wherein, in step one, gram identification index is by be checked
Asking in the gram set of character string, the summation of gram number with target gram with identical MinHash signature represents, described always
Do not comprise described target gram.
The present invention proposes a kind of similar character string search method, wherein, with the actually used table of falling row chain in step 2
The number of element calculates the table of falling row chain length, and is added by ratio based on chained list length by weight and integrated described gram and distinguish
Knowledge and magnanimity index and the table of falling row chain length index, form aggregative indicator.
The present invention proposes a kind of similar character string search method, and wherein, described step 3 is by described extensive prefix mistake
Filter obtains retrieval result and comprises the steps:
Step A1: after inquiry string is split by gram, the polling character after splitting by described aggregative indicator
String carries out ascending sort；
Step A2: choosing+1 gram of front τ of the non-overlapping part of character string as base gram query set, wherein τ is distance
Threshold values；
Step A3: new gram the gain after calculating increase are added in circulation, until newly-increased gram do not have gain or
Till there is no new gram；
Step A4: all gram are merged, and take out occurrence number be at least LB time character string add Candidate Set,
Character string in described Candidate Set is carried out real editing distance calculating, thus obtains retrieval result；Wherein, LB refers to described
The lower limit of the identical gram number that extensive prefix filter uses.
The present invention proposes a kind of similar character string search method, and wherein, the gain in described step A3 is with following formula
Represent:
In formula, G represents gain, CiRepresent the size of current Candidate Set, | I (gn) | represent gnThe table of falling row chain length, Js
Representing Jaccard similarity, cost (Q) represents that the average time that inquiry carries out once true editing distance calculating is time-consuming.
The present invention proposes extensive prefix filter, and utilizes the coincidence information between the table of falling row chain to propose new weighing apparatus
The index of amount gram mass.The present invention is by the extensive prefix filter after optimizing, it is possible to increase recall precision, especially works as volume
Collect when threshold values uprises, very big performance boost can be obtained.Present invention also overcomes traditional method, it is impossible to adaptively selected
The shortcoming of excellent filter.
Accompanying drawing explanation
Fig. 1 represents the flow chart of similar character string search method of the present invention；
Fig. 2 represents the inverted index in embodiment.
Fig. 3 represents the MinHash signature in embodiment.
Detailed description of the invention
In conjunction with specific examples below and accompanying drawing, the present invention is described in further detail.Implement the present invention process,
Condition, experimental technique etc., outside the lower content mentioned specially, be universal knowledege and the common knowledge of this area, this
Bright content is not particularly limited.
The similar character string search method of the present invention introduces a kind of new extensive prefix filter, it is achieved subtract to greatest extent
In few Candidate Set.Similar character string search method of the present invention makes full use of the information of inverted index, propose to utilize the table of falling row chain it
Between overlay information as index, in the case of not reducing performance, increase high-quality gram to reduce Candidate Set.Although ratio
Prior art adds estimation expense, but greatly improves retrieval performance.As it is shown in figure 1, similar character string of the present invention retrieval
Method comprises the following steps:
Step one: utilize the coincidence information between the table of falling row chain in inverted index, obtains gram identification index；Wherein,
During gram identification index is gathered by the gram of character string to be checked in step one, with target gram, there is identical MinHash and sign
The summation of the gram number of name represents, does not comprise target gram in this summation.
Further, first to needing the string assemble being retrieved to set up inverted index before step one, and generate
The MinHash signature that inverted index is corresponding.The present invention splits with q-gram, after splitting by being numbered character string
Gram is entry maps sets up inverted index to the relation of the character string comprising gram.Length according to character string chooses Hash letter
Number, generates MinHash signature in conjunction with inverted index.
Step 2: gram identification index integrated with the table of falling row chain length ratio, forms aggregative indicator.Wherein,
Step 2 calculates the table of falling row chain length with the number of the actually used table element of falling row chain, and by ratio based on chained list length
Example is added by weight and integrates described gram identification index and the table of falling row chain length index, forms aggregative indicator.
Step 3: obtain Candidate Set by extensive prefix filter according to aggregative indicator, calculates the character string in Candidate Set
Real editing distance obtains retrieval result.Wherein, step 3 includes following by extensive prefix filter acquisition retrieval result
Step:
Step A1: after splitting inquiry string by gram, the inquiry string after splitting by aggregative indicator enters
Row ascending sort；
Step A2: choosing+1 gram of front τ of the non-overlapping part of character string as base gram query set, wherein τ is distance
Threshold values；
Step A3: new gram the gain after calculating increase are added in circulation, until newly-increased gram do not have gain or
Till there is no new gram；Wherein, the gain in step A3 represents with following formula:
In formula, G represents gain, CiRepresent the size of current Candidate Set, | I (gn) | represent gnThe table of falling row chain length, Js
Representing Jaccard similarity, cost (Q) represents that the average time that inquiry carries out once true editing distance calculating is time-consuming.
Step A4: all gram are merged, and take out occurrence number be at least LB time character string add Candidate Set,
Character string in Candidate Set carrying out real editing distance calculating, thus obtains retrieval result, wherein LB refers to extensive prefix
The lower limit of the identical gram number that filter uses.
In the present invention, extensive prefix filter refer to character string carrying out similarity query when, by aggregative indicator
In gram set after ascending sort, the only some gram in foremost with sequence filter the filtration side of incoherent character string
Formula.
Embodiment:
In the present invention, two parts have been needed at pretreatment stage: set up inverted index and produce MinHash label
Name.
Inverted index needs record mapping relations from gram to character string.Table 1 is a string assemble, comprises 5
Character string, and use 0～4 to be numbered.
The character string of table 1 the present embodiment
| id | Strings |
| 0 | rich |
| 1 | stick |
| 2 | stich |
| 3 | stuck |
| 4 | stuch |
First all 5 character strings are carried out 2-gram fractionation, then sets up, as entrance, the row of falling using the gram occurred
Index, it is thus achieved that inverted index as shown in Figure 2.Wherein, the table of falling row chain preserves is to comprise current gram in string assemble
All character strings.
Owing to the generation process of MinHash signature needs replacement operator, hash function is generally used to simulate replacement operator.
Table 2 is two hash functions.The character string of equal length is used same group of hash function.The present embodiment is often organized hash function
The most only comprise a hash function.
The hash function that the different group of table 2 uses
| String length | Hash function |
| 4 | (x+1) %2 |
| 5 | (2x+3) %5 |
First hash function is used in the character string of numbering 0 because its a length of 4, second hash function is used in
The character string of numbering 1～4, because their length is all 5.As a example by ch, inside length 4 interval, because only that a word
Symbol string, its MinHash signature value is (0+1) %2=1.Inside length 5 interval, there are 4 character strings, calculate every the most respectively
The non-zero value of the individual character string first appearance after being upset by hash function.Because the numbering in the table of falling row chain i.e. represents
It is the position of 1 after vectorization, therefore directly the numbering in chained list is substituted into successively hash function, and selects the position of minimum.
Second hash function is used to obtain (2*2+3) %5=2 and (2*4+3) %5=1 two non-zero positions, because a rear minimum,
So selecting the cryptographic Hash of latter to sign as the MinHash of length of interval 5.MinHash signature last for ch is 1-1.Fig. 3
For the MinHash signature generated according to the inverted index in Fig. 2.
For the quality of gram, the present invention proposes a kind of new aggregative indicator: length based on the gram table of falling row chain and
The identification of gram.The identification index of gram is used to weigh gram and produces the degree that candidate result is how many, identification index
The lowest, gram is less susceptible to together with other gram produce candidate character strings, and therefore the quality of gram is the highest, it should than other
Gram prioritizing selection.Owing to selecting any D+LB gram to ensure in extensive gram filter, result is correct, but different
Gram combination produce filter capability have the biggest difference, the identification of gram is just used to the quantization of assisted Selection gram and refers to
Mark, represents the quality of gram.Gram identification is a relative indicatrix, it is necessary to be combined into base with a concrete q-gram collection
Plinth, it is defined as follows: the q-gram for given inquiry string S gathers Gq (S), for any one gram gi
∈ Gq (S), I (gi) represent in inverted index with gram giFor the set of all character strings of the table of falling row chain of entrance, for
Arbitrarily ej∈I(gi), first count in Gq (S) and the table of falling row chain of each gram comprises ejGram number, and this value is deducted 1,
It is designated as the identification of each element, I (gi) identification of each element in the table of falling row chain sums up i.e. acquisition gram gi
Identification.But owing to the table of falling row chain is generally the longest, this index of Practical Calculation is the most time-consuming in queries, therefore this
Bright by estimating gram identification based on MinHash signature, its method of estimation is similar with calculating true identification, only need to will calculate
The method of identification changes MinHash into from original row and signs.As, making current inquiry is strick, and τ=1,
The 2-gram collection of its inquiry is combined into { st, tr, ri, ic, ck}.As a example by st, due to τ=1, string length L of strick is
6, according to the character of string search, only need to consider in the table of falling row chain, in the part of interval [L-τ, L+ τ], i.e. only interval 5
MinHash signature needs to consider.In length of interval 5, the MinHash signature of st is 0, at the 2-that all inquiries split out
In gram, sign under same hash function be also 0 have ck, ic and st3, therefore the estimation identification of st is 3-1=2.
The table of falling row chain length is merged with identification and generates total weight of gram and directly can respectively take the weight of 1/2 after regularization
Add and.Table 3 representing, merging weight represents each gram gram mass in each interval.
Table 3gram identification and the weight of the table of falling row chain length ratio
| gram | Identification 4 | Identification 5 | Length 5 | Merge weight 5 | Total weight |
| st | 0 | 2 | 4 | 0.42 | 0.42 |
| tr | null | null | null | null | null |
| ri | 1 | 0 | 0 | 0 | 0 |
| ic | 1 | 2 | 2 | 0.29 | 0.29 |
| ck | 0 | 2 | 2 | 0.29 | 0.29 |
When merging the weight in different interval, use and add in the ratio of the different interval character string numbers comprised
With.Therefore, for st, total identification is exactly 0*0.2+0.42*0.8=0.336.
Present invention search framework based on extensive prefix filter is divided into two steps: generate base inquiry gram set new with interpolation
Gram.First the gram of query generation is carried out ascending sort according to the comprehensive quality index of gram before carrying out before execution.
As shown in the final weight of table 3, final order is ri, ic, ck, st.Due to have between the gram including same position together with
Time affected in this total position edit operation by occurring, if newly-increased gram comprises the position selected, will not bring
Performance increases, therefore, selecting gram when, if after gram select above, newly-increased gram can not with select
Gram sharing position.Such as character string abcdef, after gram ab, then bc cannot use because ab with
Bc have shared b.After sequence, select to meet front minimum+1 gram of τ without sharing position as base inquiry gram set.This reality
Execute example and select ri and ck (because ri Yu ic has a shared part, all do not select ic), merge the table of falling row chain of the two gram, own
Occur in the element 1 and 3 on chained list all as candidate result.The step adding new gram is a cyclic process, when not having property
The when of energy gain, circulation stops.WithRepresent gain.Wherein, CiRepresent current
The size of Candidate Set, | I (gn) | represent gnThe table of falling row chain length, JsRepresenting Jaccard similarity, cost (Q) represents looking into
The average time that inquiry carries out once true editing distance calculating is time-consuming.If this gain is more than 1, then mean that current newly-increased
Gram can improve performance, can add in Candidate Set.Newly-increased gram is the most down found according to sequence before, the most right
St carries out gain assessment.By calculating, increasing st does not has performance gain, therefore cannot add Candidate Set, and circulation stops.This reality
Execute in example Candidate Set for 1,3}, be stick, stuck}, the two candidate result and strick are truly edited away from
After calculating, the only true editing distance of character string 1 meets with inquiry strick distance within 1, so obtaining character
String 1 is the result meeting condition.
The protection content of the present invention is not limited to above example.Under the spirit and scope without departing substantially from inventive concept, this
Skilled person it is conceivable that change and advantage be all included in the present invention, and with appending claims for protect
Protect scope.
Claims (4)
1. a similar character string search method, it is characterised in that comprise the following steps:
Step one: utilize the coincidence information between the table of falling row chain in inverted index, obtains gram identification index；Wherein, described
Gram identification index is by the gram set of character string to be checked, has identical MinHash signature with target gram
The summation of gram number represents, does not comprise target gram in this summation described；
Step 2: described gram identification index integrated with the table of falling row chain length ratio, forms aggregative indicator；Wherein,
The table of falling row chain length is calculated with the number of the actually used table element of falling row chain, and by ratio based on chained list length by weight
Add and integrate described gram identification index and the table of falling row chain length index, forming aggregative indicator；
Step 3: obtain Candidate Set by extensive prefix filter according to described aggregative indicator, by calculating in described Candidate Set
The real editing distance of character string obtain retrieval result；Wherein, retrieval result bag is obtained by described extensive prefix filter
Include following step:
Step A1: after splitting inquiry string by gram, the inquiry string after splitting by described aggregative indicator enters
Row ascending sort；
Step A2: choose+1 gram of front τ of the non-overlapping part of character string as base gram query set, wherein τ is distance threshold values；
Step A3: circulate the gain after adding new gram and calculating increase, until newly-increased gram does not has gain or does not has
Till new gram；Wherein, described gain represents with following formula:
In formula, G represents gain, CiRepresent the size of current Candidate Set, | I (gn) | represent gnThe table of falling row chain length, gn represents
Gram after the fractionation of sequence number n, JsRepresent that Jaccard similarity, cost (Q) expression carry out once true editing distance to inquiry
The average time calculated is time-consuming；
Step A4: all gram are merged, and take out occurrence number be at least LB time character string add Candidate Set, to institute
State the character string in Candidate Set and carry out real editing distance calculating, thus obtain retrieval result；Wherein, LB refers to described extensive
The lower limit of the identical gram number that prefix filter uses.
Similar character string search method the most according to claim 1, it is characterised in that before step one, first to needs
The string assemble being retrieved sets up inverted index, and generates the MinHash signature that inverted index is corresponding.
Similar character string search method the most according to claim 2, it is characterised in that by character string is numbered with
Q-gram splits, and the gram after splitting sets up inverted index for entry maps to the relation of the character string comprising described gram.
Similar character string search method the most according to claim 2, it is characterised in that select according to the length of described character string
Take hash function, generate MinHash signature in conjunction with described inverted index.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310294529.4A CN103365998B (en) | 2013-07-12 | 2013-07-12 | A kind of similar character string search method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310294529.4A CN103365998B (en) | 2013-07-12 | 2013-07-12 | A kind of similar character string search method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN103365998A CN103365998A (en) | 2013-10-23 |
| CN103365998B true CN103365998B (en) | 2016-08-24 |
Family
ID=49367339
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201310294529.4A Active CN103365998B (en) | 2013-07-12 | 2013-07-12 | A kind of similar character string search method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103365998B (en) |
Families Citing this family (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105426759A (en) * | 2015-10-30 | 2016-03-23 | 百度在线网络技术（北京）有限公司 | URL legality determining method and apparatus |
| CN107153652B (en) * | 2016-03-03 | 2020-10-30 | 创新先进技术有限公司 | Method and device for converting target character string into normalized character string |
| CN105913094B (en) * | 2016-05-03 | 2019-06-21 | 中国科学院信息工程研究所 | A kind of minimum range character string calculating lookup method |
| CN106897990B (en) * | 2016-08-31 | 2019-10-25 | 广东工业大学 | The character defect inspection method of tire-mold |
| CN108255836B (en) * | 2016-12-28 | 2020-12-25 | 普天信息技术有限公司 | Character string matching method and device |
| CN108984695B (en) * | 2018-07-04 | 2021-04-06 | 科大讯飞股份有限公司 | Character string matching method and device |
| CN110008383B (en) * | 2019-04-11 | 2021-07-27 | 北京安护环宇科技有限公司 | Black and white list retrieval method and device based on multiple indexes |
| CN110502629B (en) * | 2019-08-27 | 2020-09-11 | 桂林电子科技大学 | LSH-based connection method for filtering and verifying similarity of character strings |
| CN110533035B (en) * | 2019-08-28 | 2022-02-15 | 海南阿凡题科技有限公司 | Student homework page number identification method based on text matching |
| CN111078821B (en) * | 2019-11-27 | 2023-12-08 | 泰康保险集团股份有限公司 | Dictionary setting method, dictionary setting device, medium and electronic equipment |
| CN112486989B (en) * | 2020-11-28 | 2021-08-27 | 河北省科学技术情报研究院(河北省科技创新战略研究院) | Multi-source data granulation fusion and index classification and layering processing method |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101763405A (en) * | 2009-11-16 | 2010-06-30 | 陆嘉恒 | Approximate character string searching technology based on synonym rule |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8271499B2 (en) * | 2009-06-10 | 2012-09-18 | At&T Intellectual Property I, L.P. | Incremental maintenance of inverted indexes for approximate string matching |
-
2013
- 2013-07-12 CN CN201310294529.4A patent/CN103365998B/en active Active
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101763405A (en) * | 2009-11-16 | 2010-06-30 | 陆嘉恒 | Approximate character string searching technology based on synonym rule |
Non-Patent Citations (3)
| Title |
|---|
| Can We Beat the Prefix Filtering?An Adaptive Framework for Similarity Join and Search;Jiannan Wang等;《Proc. ACM SIGMOD Int’l Conf. Management of Data (SIGMOD)》;20120524;第85-96页 * |
| Efficient Exact Edit Similarity Query Processing with the Asymmetric Signature Scheme;Jianbin Qin等;《Proc. ACM SIGMOD Int’l Conf. Management of Data(SIGMOD)》;20110616;第1033-1044页 * |
| VChunkJoin:An Efficient Algorithm for Edit Similarity Joins;Wei Wang等;《IEEE Tran.on Knowledge and Data Engineering》;20120417;第1916-1929页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN103365998A (en) | 2013-10-23 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103365998B (en) | A kind of similar character string search method | |
| CN104252445B (en) | Approximate repetitive file detection method and device | |
| CN103678576B (en) | The text retrieval system analyzed based on dynamic semantics | |
| US20150356128A1 (en) | Index key generating device, index key generating method, and search method | |
| CN107590128B (en) | Paper homonymy author disambiguation method based on high-confidence characteristic attribute hierarchical clustering method | |
| CN109117464A (en) | A kind of data similarity detection method based on editing distance | |
| CN106649597A (en) | Method for automatically establishing back-of-book indexes of book based on book contents | |
| CN103177120B (en) | A kind of XPath query pattern tree matching method based on index | |
| CN111127068B (en) | Automatic pricing method and device for engineering quantity list | |
| CN107291895B (en) | Quick hierarchical document query method | |
| CN106445906A (en) | Generation method and apparatus for medium-and-long phrase in domain lexicon | |
| CN110096699A (en) | Semantic-based machine reads the candidate answers screening technique understood and system | |
| CN105488212A (en) | Data quality detection method and device of duplicated data | |
| CN102693320B (en) | Searching method and device | |
| JP7497403B2 (en) | Method and system for performing product matching on an electronic commerce platform - Patents.com | |
| CN103425711B (en) | Object value alignment schemes based on many object instances | |
| CN106155998A (en) | A kind of data processing method and device | |
| CN115794833A (en) | Data processing method, server and computer storage medium | |
| CN101256594A (en) | Method and system for measuring graph structure similarity | |
| CN112148735B (en) | Construction method for structured form data knowledge graph | |
| Chu et al. | Automatic data extraction of websites using data path matching and alignment | |
| JPWO2009048149A1 (en) | Electronic document equivalence judgment system and equivalence judgment method | |
| CN103970732B (en) | Mining method and device of new word translation | |
| CN111325235A (en) | Multilingual-oriented universal place name semantic similarity calculation method and application thereof | |
| CN108920134B (en) | Method and device for automatically generating design document |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant |