# data-obfuscator KEYMAP2 - PRIVATE. Never share this file. # It turns the masked ADaM datasets (and the NONMEM-style dataset an LLM builds from them) back into # real values: `data-obfuscator reverse --input master.csv --keymap keymap2.txt`. # # Records (TAB-separated; the first field names the record): # file # domain_column a column (in any file) that belongs to the identifier domain # recode one table per domain, shared by every file # date_column dates in this column were moved by the shift below # scale one factor per column name, the same in every file # not_restorable values that were redacted or dropped and cannot come back # format keymap2 1 program 0.0.0.23 seed 7 level 3 date_shift_days 3 date_shift_whole_days 3 # 1. FILES file 1 adsl.csv adsl_masked.csv 12 a3f9b9d9702f8fd1d691cc5ace32a5fce23d151044a9db7d82ff18f5b983d975 file 2 adex.csv adex_masked.csv 24 5a48c98256e085462d4d1d0c4ca2d0dd162b9f0fb8c3216a6d17417f7f489aef file 3 adpc.csv adpc_masked.csv 168 65a3a211f4748cb5a97c652b0f0c8a30b9af44d3199355eb74955516391e4bb2 file 4 adlb.csv adlb_masked.csv 72 8c2d8253ced8df0579d0f47d85a010239c44f873faddd27ecab978b4e0965dbb # 2. IDENTIFIER AND LABEL RECODING (shared by all files) domain_column study STUDYID domain_column subject USUBJID domain_column subject SUBJID domain_column site SITEID domain_column user INVNAM domain_column treatment ARM domain_column treatment TRT01P domain_column treatment TRT01PN domain_column sex SEX domain_column race RACE recode study ABC-PK-2024 STU001 recode subject ABC-PK-2024-0142-1001 SUBJ001 recode subject ABC-PK-2024-0142-1002 SUBJ002 recode subject ABC-PK-2024-0142-1003 SUBJ003 recode subject ABC-PK-2024-0142-1004 SUBJ004 recode subject ABC-PK-2024-0142-1005 SUBJ005 recode subject ABC-PK-2024-0142-1006 SUBJ006 recode subject ABC-PK-2024-0267-1007 SUBJ007 recode subject ABC-PK-2024-0267-1008 SUBJ008 recode subject ABC-PK-2024-0267-1009 SUBJ009 recode subject ABC-PK-2024-0267-1010 SUBJ010 recode subject ABC-PK-2024-0267-1011 SUBJ011 recode subject ABC-PK-2024-0267-1012 SUBJ012 recode subject 1001 SUBJ013 recode subject 1002 SUBJ014 recode subject 1003 SUBJ015 recode subject 1004 SUBJ016 recode subject 1005 SUBJ017 recode subject 1006 SUBJ018 recode subject 1007 SUBJ019 recode subject 1008 SUBJ020 recode subject 1009 SUBJ021 recode subject 1010 SUBJ022 recode subject 1011 SUBJ023 recode subject 1012 SUBJ024 recode site 0142 SITE001 recode site 0267 SITE002 recode user DR KIM USER001 recode user DR PATEL USER002 recode user DR SMITH USER003 recode treatment Drug 50 mg ARM_A recode treatment Drug 100 mg ARM_B recode treatment Placebo ARM_C recode treatment 1 ARM_D recode treatment 2 ARM_E recode treatment 0 ARM_F recode sex M 0 recode sex F 1 recode race BLACK CAT001 recode race WHITE CAT002 recode race ASIAN CAT003 # 3. DATE SHIFT (days above; applies to these columns) date_column TRTSDT date_column TRTEDT date_column BRTHDT date_column EXSTDTC date_column ADTM date_column ADT # 4. NUMERIC SCALING (original = masked x factor for divide, masked / factor for multiply) scale AGE multiply 11.157546561650236 scale WEIGHTBL divide 6.392395639678136 scale AVAL divide 8.788277622606675 # 5. NOT RESTORABLE not_restorable adsl.csv COMMENT free text redacted (.)