AWS app to match two fields in two different tables with fuzzy-look up (USE Cloud Search, Glue .... others)

We have 3 data sources in xls or csv: 1st Coming from LinkedIn extractor > contains individual mails + company data 2nd From Italia Company Register > contains company data 3rd Coming from emails&events > contains individual data -ST1 These data will be collected continuosly on a weekly base and data must be stored to a DB in 3 different tables, avoiding duplicate records with primary key: Table 1 Linkedin: member_id (two records can't have same member_id) Table 2 Company Register: POB_id or VAT_id (two records can't have same POB_id or VAT_id) Table 3 Leads: individual email address (two records can't have same email address) In this 3 tables a column status must be added with initial value "To Be Qualified" The sys also must have 3 "service table" to standardize data extracted from linkedin Talble 4 (Org_Id on Linkedin, Pod_id) to match Table I and II Table 5 (organization_title_1, Role) to standardize role Table 6 (location_name, Province) to standardize province -only for Italian 1.1. STANDARDIZE ROLE The app must look up in Table1 organization_title_1 and find a match on Role 5 table for example if we have HR director> HR or Talent Developement>HR this shloud be performerd trough semantic search 1.2. STANDARDIZE LOCATION The up must look-up in location_name only for Italian and find a similar province in the table trough semantic search This operation do not needs to be supervised by the user - ST2 Once the data are uploaded the user start second step "qualify company" button to match the data across the 3 tables a 4rd table is created with the following fields - linkedin member_id (primary key) - Name Surname (from Linkedin) - Email from Linkedin - Country from Linkedin - Industry form Linkedin - Company from Linkedin > 2.1 first of all the sys check in Table 4 if there is already a match between organization_id_1 (on table 1 LdExtractor) and Company Table PoBId. If there is a Match go to 2.3. if not continue 2.2. 2.2. the app must look into the second table (Italian Company Register) with fuzzy look-up to find an approximated matches for company name (i.e. Comapny on <<Linkedin Barilla>> Company on TableII Registry <<Flli. Barilla Spa>> or just <<[login to view URL] Barilla>>) The sys must show a set of possible matches between company name in Linkedin Extractor (Table 1) and in company registry (Table 2) the user must select one. If a match is found and the organization_id_1 (on table 1 LdExtractor) is not emplty then all records with same id may be matched and qualified. For example Accenture=1033 all Accenture employee will share same organization identifier. Please note this choiche must be a supervised process. 2.3. If a match is found Linkedin table1 status="Qualified" and Italian Company registry="Qualified" If no match is found Linkedin table Status:="In qualification" and Italian Company registry="In qualification" (A list of all company "In qualification" will be qualified manually later) NOTE: This fuzzy search may be performed with AWS cloudsearch or glue or similar tools Then all company register fields must be merged - Pod_Id, Adress City, Province etc ..... activities.... - ST3 third step will be "Qualify individual" the app must look into the 3rd table and look for mail or (Name, Surname) for exact match (no case sensitive). This table contains one or more email address and phone numbers of the individual. The sys must show a set of possible matches between Table1 Name and Surname and table 3 Name and Surname as we may have cases of homonymy. The user must choose the exact match. If match is found record on Table 3 status ="Qualified" Then all 3rd table fields must be merged - mail phone etc the primary key of the new table 4rd will be linkedin member_id (primary key) one individual may have more emails but just onw member_id A list of record will be sent by chat