chore: 레지스터맵 생성 스크립트 정리 + field_hist import

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

scripts/build_register_map.py

@@ -2,52 +2,52 @@
 """
 HC900 Register Map Builder
 
-Converts HC Designer CSV exports (SummaryFunctionBlockReport, SignalTags, Variables)
+Converts Sinam_Tag_all.xlsx (Experion point database) as the PRIMARY source
-into a unified JSON register map for the Modbus TCP gateway.
+for PID loop tag names, with HC Designer CSV exports for address validation.
 
 Usage:
     python3 build_register_map.py \
-        --loop-csv ../docs/SummaryFucntionBlockReport.csv \
+        --xlsx ../docs/Sinam_Tag_all.xlsx \
         --signal-csv ../docs/SignalTags.csv \
         --variable-csv ../docs/Variables.csv \
+        --loop-csv ../docs/SummaryFucntionBlockReport.csv \
         -o ../docs/register-map.json
 """
 
 import csv
 import json
 import argparse
+import re
 from pathlib import Path
 
+try:
+    import openpyxl
+except ImportError:
+    print("ERROR: openpyxl is required. Install with: pip install openpyxl")
+    exit(1)
+
 # ─── PID loop parameter offset map (based on manual Table 6-3) ───
 
 LOOP_PARAM_OFFSETS = [
-    # (name, offset_hex, type, access, description)
-    # PV / SP / OP
     ("PV",               0x00, "float32", "R",    "Process Variable"),
     ("RSP",              0x02, "float32", "R",    "Remote Set Point (SP2)"),
     ("SP",               0x04, "float32", "RW",   "Working Set Point"),
     ("OP",               0x06, "float32", "RW",   "Output"),
-    # Gains / tuning
     ("GAIN1",            0x0C, "float32", "RW",   "Gain #1 / Prop Band #1"),
     ("DIR",              0x0E, "float32", "R",    "Direction (0=Direct, 1=Reverse)"),
     ("RESET1",           0x10, "float32", "RW",   "Reset #1"),
     ("RATE1",            0x12, "float32", "RW",   "Rate #1"),
-    # Ranges
     ("PV_LO",            0x16, "float32", "R",    "PV Low Range"),
     ("PV_HI",            0x18, "float32", "R",    "PV High Range"),
-    # Alarms
     ("ALM1_SP1",         0x1A, "float32", "RW",   "Alarm #1 SP #1"),
     ("ALM1_SP2",         0x1C, "float32", "RW",   "Alarm #1 SP #2"),
-    # Local SPs
     ("LSP1",             0x2A, "float32", "RW",   "Local SP #1"),
     ("LSP2",             0x2C, "float32", "RW",   "Local SP #2"),
-    # Limits
     ("SP_LO",            0x34, "float32", "RW",   "SP Low Limit"),
     ("SP_HI",            0x36, "float32", "RW",   "SP High Limit"),
     ("OP_LO",            0x3A, "float32", "RW",   "Output Low Limit"),
     ("OP_HI",            0x3C, "float32", "RW",   "Output High Limit"),
     ("DEV",              0x4A, "float32", "R",    "Deviation (SP-PV)"),
-    # Bit-packed status / control
     ("FUZZY_EN",         0xB7, "uint16",  "RW",   "Fuzzy Enable"),
     ("ATUNE_REQ",        0xB8, "uint16",  "RW",   "Autotune Request"),
     ("MODE",             0xBA, "uint16",  "RW",   "Auto/Manual (0=Man, 1=Auto)"),
@@ -56,11 +56,35 @@ LOOP_PARAM_OFFSETS = [
     ("LOOP_STATUS",      0xBE, "uint16",  "R",    "Loop Status (bit packed)"),
 ]
 
+# Experion source param → canonical param name mapping
+SOURCE_PARAM_MAP = {
+    "PV":       "PV",
+    "WSP":      "SP",
+    "SPWORK":   "SP",
+    "OPWORK":   "OP",
+    "OP":       "OP",
+    "GAIN1":    "GAIN1",
+    "RESET1":   "RESET1",
+    "RATE1":    "RATE1",
+    "MODEIN":   "MODE",
+    "DIRECTION": "DIR",
+    "LOOPSTAT": "LOOP_STATUS",
+}
+
+# Source columns in Sinam_Tag_all.xlsx (0-indexed) where LOOP params appear
+LOOP_SOURCE_COLS = [26, 27, 28, 29, 31, 32, 33, 75]
+
 # ─── helpers ───
 
 def hex_addr(s: str) -> int:
     return int(s.strip(), 16)
 
+
+def normalize_tag(tag: str) -> str:
+    """Normalize tag name: remove hyphens, uppercase."""
+    return tag.replace("-", "").upper()
+
+
 def strip_header(rows, header_marker: str):
     """Skip rows until we find the actual data header."""
     for i, row in enumerate(rows):
@@ -68,7 +92,25 @@ def strip_header(rows, header_marker: str):
             return rows[i + 1:]
     return rows
 
-# ─── parsers ───
+
+# ─── Experion point naming: append .PV for standard instrument tags ───
+
+_EXPERION_PV_PREFIXES = frozenset([
+    "DP", "FI", "FIQ", "FQ", "FT", "LI", "LIA", "LIC", "LICA", "LIS",
+    "LISA", "PI", "PIA", "PIC", "PICA", "PT", "TI", "TIA", "TIC", "TICA",
+    "TE", "WT",
+])
+
+
+def _to_experion_point_name(tag: str) -> str:
+    """Append .PV for standard instrument tags per Experion naming convention."""
+    m = re.match(r"^([A-Z]+)", tag)
+    if m and m.group(1) in _EXPERION_PV_PREFIXES:
+        return tag + ".PV"
+    return tag
+
+
+# ─── CSV parsers (unchanged from original) ───
 
 def parse_loop_csv(path: Path) -> list[dict]:
     """Parse SummaryFunctionBlockReport.csv → list of loop dicts."""
@@ -76,7 +118,6 @@ def parse_loop_csv(path: Path) -> list[dict]:
     with open(path, "r", encoding="utf-8-sig") as f:
         reader = csv.reader(f)
         rows = list(reader)
-
     for row in rows:
         if len(row) < 6:
             continue
@@ -105,7 +146,6 @@ def parse_signal_or_variable_csv(path: Path, access: str) -> list[dict]:
     with open(path, "r", encoding="utf-8-sig") as f:
         reader = csv.reader(f)
         rows = list(reader)
-
     for row in rows:
         if len(row) < 8:
             continue
@@ -116,7 +156,6 @@ def parse_signal_or_variable_csv(path: Path, access: str) -> list[dict]:
         desc = row[3].strip()
         typ = row[6].strip() if len(row) > 6 else "float 32"
         eu = row[7].strip() if len(row) > 7 else ""
-        # normalise type
         if "float" in typ.lower():
             data_type = "float32"
             register_count = 2
@@ -126,10 +165,9 @@ def parse_signal_or_variable_csv(path: Path, access: str) -> list[dict]:
         else:
             data_type = "float32"
             register_count = 2
-
         addr = hex_addr(hex_addr_raw)
         tags.append({
-            "tag": tag,
+            "tag": _to_experion_point_name(tag),
             "description": desc,
             "addr": addr,
             "count": register_count,
@@ -140,34 +178,198 @@ def parse_signal_or_variable_csv(path: Path, access: str) -> list[dict]:
     return tags
 
 
+# ─── PID loop tag name matching ───
+
+def _make_csv_lookup(loops: list[dict]) -> dict:
+    """
+    Build lookup from normalized tag name → (base_addr, csv_tag, loop_number).
+    Handles FICA↔FICQ prefix variants.
+    """
+    lookup = {}
+    for loop in loops:
+        norm = normalize_tag(loop["tag"])
+        # Primary entry
+        lookup[norm] = (loop["start_addr"], loop["tag"], loop["loop_number"])
+        # Also add the FICA↔FICQ variant
+        if norm.startswith("FICA"):
+            alt = "FICQ" + norm[4:]
+            if alt not in lookup:
+                lookup[alt] = (loop["start_addr"], loop["tag"], loop["loop_number"])
+        elif norm.startswith("FICQ"):
+            alt = "FICA" + norm[4:]
+            if alt not in lookup:
+                lookup[alt] = (loop["start_addr"], loop["tag"], loop["loop_number"])
+    return lookup
+
+
+def _resolve_loop_addr(xlsx_tag: str, csv_lookup: dict) -> tuple:
+    """
+    Try to find the address for an xlsx tag from the CSV lookup.
+    Returns (addr, csv_tag, loop_number) or (None, None, None).
+    """
+    norm = normalize_tag(xlsx_tag)
+    if norm in csv_lookup:
+        return csv_lookup[norm]
+    return (None, None, None)
+
+
+# ─── xlsx parsing ───
+
+def parse_xlsx_loops(path: Path, csv_lookup: dict) -> list[dict]:
+    """
+    Parse Sinam_Tag_all.xlsx for PID loop AnalogPoint entries.
+    Returns list of register entry dicts.
+    """
+    wb = openpyxl.load_workbook(path, data_only=True)
+    ws = wb.active
+
+    registers = []
+    warnings = []
+    validated_count = 0
+    unvalidated_count = 0
+
+    # Map param_name → (canonical_name, offset, type, access, desc_suffix)
+    param_defs = {
+        entry[0]: entry
+        for entry in LOOP_PARAM_OFFSETS
+    }
+
+    # Suffix mapping for tag names
+    PARAM_SUFFIX = {
+        "PV": ".PV",
+        "SP": ".SP",
+        "OP": ".OP",
+        "GAIN1": ".GAIN1",
+        "RESET1": ".RESET1",
+        "RATE1": ".RATE1",
+        "MODE": ".MODE",
+        "LOOP_STATUS": ".LOOPSTAT",
+    }
+
+    matched_tags = {}  # item_name → (addr, csv_tag, loop_num)
+
+    for r in range(3, ws.max_row + 1):
+        item_name = ws.cell(r, 1).value
+        cls_val = ws.cell(r, 2).value
+
+        if not item_name or cls_val != "AnalogPoint":
+            continue
+
+        # Scan source columns for LOOP references
+        loop_params = {}  # canonical_param_name → source_param_name
+        for c in LOOP_SOURCE_COLS:
+            v = ws.cell(r, c + 1).value  # openpyxl is 1-indexed
+            if v and "LOOP" in str(v):
+                m = re.match(r"C(\d+)\s+LOOP\s+(\d+)\s+(\S+)", str(v))
+                if m:
+                    c_unit = int(m.group(1))
+                    loop_num = int(m.group(2))
+                    source_param = m.group(3)
+                    canon = SOURCE_PARAM_MAP.get(source_param)
+                    if canon:
+                        loop_params[canon] = source_param
+
+        if not loop_params:
+            continue
+
+        # Find address from CSV lookup
+        addr, csv_tag, csv_loop_num = _resolve_loop_addr(item_name, csv_lookup)
+
+        # If no direct match, try the second time with the tag without suffix if it has one
+        if addr is None and "." in item_name:
+            base_name = item_name.split(".")[0]
+            addr, csv_tag, csv_loop_num = _resolve_loop_addr(base_name, csv_lookup)
+
+        if addr is not None:
+            validated = True
+            validated_count += 1
+            matched_tags[item_name] = (addr, csv_tag, csv_loop_num)
+        else:
+            validated = False
+            unvalidated_count += 1
+            warnings.append(
+                f"WARNING: No CSV match for '{item_name}' (C{c_unit} LOOP {loop_num}), "
+                f"params: {list(loop_params.keys())}"
+            )
+            continue
+
+        description_base = f"Loop #{csv_loop_num} {csv_tag}" if csv_tag else f"C{c_unit} Loop {loop_num}"
+
+        for canon, source_param in sorted(loop_params.items()):
+            if canon not in param_defs:
+                warnings.append(f"WARNING: Unknown param '{canon}' for {item_name}")
+                continue
+
+            name, off, dtype, access, desc_suffix = param_defs[canon]
+            reg_count = 2 if dtype == "float32" else 1
+            tag_suffix = PARAM_SUFFIX.get(canon, f".{canon}")
+
+            # Clean ItemName: remove any existing .PV/.SP etc. suffix
+            clean_item = item_name.split(".")[0] if "." in item_name else item_name
+            tag_name = f"{clean_item}{tag_suffix}"
+
+            description = f"{description_base} / {desc_suffix}"
+
+            registers.append({
+                "tag": tag_name,
+                "addr": addr + off,
+                "count": reg_count,
+                "type": dtype,
+                "access": access,
+                "description": description,
+                "validated": validated,
+            })
+
+    wb.close()
+    return registers, warnings, validated_count, unvalidated_count
+
+
 # ─── main ───
 
-def build(loop_path, signal_path, variable_path, output_path):
+def build(xlsx_path, loop_path, signal_path, variable_path, output_path):
-    loops = parse_loop_csv(loop_path)
     signals = parse_signal_or_variable_csv(signal_path, "R")
     variables = parse_signal_or_variable_csv(variable_path, "RW")
 
     registers = []
 
-    # 1. Expand PID loops into individual parameters
+    # Parse CSV for validation lookup
-    for loop in loops:
+    if loop_path:
-        base = loop["start_addr"]
+        loops = parse_loop_csv(loop_path)
-        for param in LOOP_PARAM_OFFSETS:
+        csv_lookup = _make_csv_lookup(loops)
-            name, off, dtype, access, desc_template = param
+    else:
-            reg_count = 2 if dtype == "float32" else 1
+        loops = []
-            addr = base + off
+        csv_lookup = {}
-            tag_name = f"{loop['tag']}.{name}"
-            description = f"{loop['description']} / {desc_template}" if loop['description'] else desc_template
-            registers.append({
-                "tag": tag_name,
-                "addr": addr,
-                "count": reg_count,
-                "type": dtype,
-                "access": access,
-                "description": description,
-            })
 
-    # 2. Signal Tags (read-only)
+    # Parse xlsx as primary PID source
+    if xlsx_path:
+        xlsx_registers, warnings, validated_count, unvalidated_count = parse_xlsx_loops(
+            xlsx_path, csv_lookup
+        )
+        registers.extend(xlsx_registers)
+    else:
+        warnings = []
+        validated_count = 0
+        unvalidated_count = 0
+        # Fallback: use CSV as loop source (original behavior)
+        for loop in loops:
+            base = loop["start_addr"]
+            for param in LOOP_PARAM_OFFSETS:
+                name, off, dtype, access, desc_template = param
+                reg_count = 2 if dtype == "float32" else 1
+                addr = base + off
+                tag_name = f"{loop['tag']}.{name}"
+                description = f"{loop['description']} / {desc_template}" if loop['description'] else desc_template
+                registers.append({
+                    "tag": tag_name,
+                    "addr": addr,
+                    "count": reg_count,
+                    "type": dtype,
+                    "access": access,
+                    "description": description,
+                })
+            validated_count = len(loops)
+
+    # Signal Tags (read-only)
     for sig in signals:
         registers.append({
             "tag": sig["tag"],
@@ -179,7 +381,7 @@ def build(loop_path, signal_path, variable_path, output_path):
             "eu": sig["eu"],
         })
 
-    # 3. Variables (R/W)
+    # Variables (R/W)
     for var in variables:
         registers.append({
             "tag": var["tag"],
@@ -191,16 +393,18 @@ def build(loop_path, signal_path, variable_path, output_path):
             "eu": var["eu"],
         })
 
-    # sort by address
+    # Sort by address
     registers.sort(key=lambda r: r["addr"])
 
     output = {
         "controller": "HC900-C70 Rev 4.4x",
-        "report_generated": "2026-06-01",
+        "report_generated": "2026-06-08",
         "float_format": "FP_B",
         "notes": "FP_B = IEEE 754 Big Endian (byte order 4,3,2,1). "
                  "Address is the first register (0-based). "
-                 "float32 uses 2 consecutive registers.",
+                 "float32 uses 2 consecutive registers. "
+                 "PID loop sources from Sinam_Tag_all.xlsx (Experion), "
+                 "addresses validated against SummaryFunctionBlockReport.csv.",
         "register_count": len(registers),
         "registers": registers,
     }
@@ -209,17 +413,26 @@ def build(loop_path, signal_path, variable_path, output_path):
         json.dump(output, f, indent=2, ensure_ascii=False)
 
     print(f"✓ Register map written to {output_path}")
-    print(f"  Loops:     {len(loops)}")
+    print(f"  PID loops (validated): {validated_count}")
-    print(f"  Signals:   {len(signals)}")
+    print(f"  PID loops (unmatched): {unvalidated_count}")
-    print(f"  Variables: {len(variables)}")
+    print(f"  Signals:               {len(signals)}")
-    print(f"  Total registers in map: {len(registers)}")
+    print(f"  Variables:             {len(variables)}")
+    print(f"  Total registers:       {len(registers)}")
+
+    for w in warnings:
+        print(f"  {w}")
+
+    return registers
 
 
 if __name__ == "__main__":
-    parser = argparse.ArgumentParser(description="Build HC900 register map JSON from CSV exports")
+    parser = argparse.ArgumentParser(
-    parser.add_argument("--loop-csv", required=True, help="SummaryFunctionBlockReport.csv")
+        description="Build HC900 register map JSON from xlsx + CSV exports"
+    )
+    parser.add_argument("--xlsx", help="Sinam_Tag_all.xlsx (primary PID loop source)")
+    parser.add_argument("--loop-csv", help="SummaryFunctionBlockReport.csv (validation)")
     parser.add_argument("--signal-csv", required=True, help="SignalTags.csv")
     parser.add_argument("--variable-csv", required=True, help="Variables.csv")
     parser.add_argument("-o", "--output", default="register-map.json", help="Output JSON path")
     args = parser.parse_args()
-    build(args.loop_csv, args.signal_csv, args.variable_csv, args.output)
+    build(args.xlsx, args.loop_csv, args.signal_csv, args.variable_csv, args.output)

scripts/build_register_map_from_csv.py

@@ -10,8 +10,24 @@ JSON register map consumed by the hc900_gateway.
 import csv
 import json
 import argparse
+import re
 from pathlib import Path
 
+# Experion point naming convention: standard instrument tags get .PV suffix.
+_EXPERION_PV_PREFIXES = frozenset([
+    "DP", "FI", "FIQ", "FQ", "FT", "LI", "LIA", "LIC", "LICA", "LIS",
+    "LISA", "PI", "PIA", "PIC", "PICA", "PT", "TI", "TIA", "TIC", "TICA",
+    "TE", "WT",
+])
+
+
+def _to_experion_point_name(tag: str) -> str:
+    """Append .PV for standard instrument tags per Experion naming convention."""
+    m = re.match(r"^([A-Z]+)", tag)
+    if m and m.group(1) in _EXPERION_PV_PREFIXES:
+        return tag + ".PV"
+    return tag
+
 
 def parse_full_map(path: Path) -> dict:
     registers = []
@@ -56,8 +72,9 @@ def parse_full_map(path: Path) -> dict:
             else:
                 dtype = "float32"
                 count = 2
+            experion_tag = _to_experion_point_name(tag)
             registers.append({
-                "tag": tag, "addr": addr, "count": count,
+                "tag": experion_tag, "addr": addr, "count": count,
                 "type": dtype, "access": "R", "description": desc,
             })
 
@@ -69,8 +86,9 @@ def parse_full_map(path: Path) -> dict:
             else:
                 dtype = "float32"
                 count = 2
+            experion_tag = _to_experion_point_name(tag)
             registers.append({
-                "tag": tag, "addr": addr, "count": count,
+                "tag": experion_tag, "addr": addr, "count": count,
                 "type": dtype, "access": "RW", "description": desc,
             })
 

scripts/import_field_hist_C3_C4.py

@@ -0,0 +1,223 @@
+#!/usr/bin/env python3
+"""
+Import field history data from field_hist DB into hc900.history_table
+for C3 and C4 controller tags only.  2026-02-05 ~ 2026-06-05 (~30s interval).
+
+Usage:
+    # Preview-only
+    python3 scripts/import_field_hist_C3_C4.py --dry-run
+
+    # Full import
+    python3 scripts/import_field_hist_C3_C4.py
+
+    # Partial date range
+    python3 scripts/import_field_hist_C3_C4.py --from-date 2026-03-01 --to-date 2026-03-07
+
+Logic:
+  1. Load C3/C4 base tag names from `register-map-c3.json` / `register-map-c4.json`.
+  2. Decode field_hist ptlist+mapping+tblist → {tagname_upper: (tblname, oit)}.
+  3. Filter to C3/C4 (by base name match), tag names stored UPPERCASE.
+  4. For each cont table in daily chunks: DELETE existing history overlap,
+     then COPY new long-format data.
+"""
+
+import argparse
+import json
+import time
+from collections import defaultdict
+from datetime import datetime, timedelta, timezone
+from pathlib import Path
+
+import psycopg
+
+KST = timezone(timedelta(hours=9))
+REG_MAP_DIR = Path(__file__).resolve().parent.parent / "docs"
+
+FIELD_DSN = "host=localhost port=5432 dbname=field_hist user=postgres password=postgres"
+TARGET_DSN = (
+    "host=localhost port=5432 dbname=iiot_platform "
+    "user=postgres password=postgres options=-csearch_path=hc900"
+)
+
+BATCH_DAYS = 1
+
+
+def load_controller_bases() -> dict[str, str]:
+    """Return {base_name_lower: controller_id} from register-map-c3/c4.json."""
+    result = {}
+    for ctrl in ("C3", "C4"):
+        path = REG_MAP_DIR / f"register-map-{ctrl.lower()}.json"
+        data = json.loads(path.read_text())
+        for entry in data["registers"]:
+            base = entry["tag"].split(".")[0].lower()
+            if base not in result:
+                result[base] = ctrl
+    return result
+
+
+def build_tag_map(allowed_bases: dict[str, str]) -> dict[str, tuple[str, int, str]]:
+    """Return {tagname_upper: (tblname, oit, controller_id)} for C3/C4 tags."""
+    tag_map = {}
+    with psycopg.connect(FIELD_DSN) as conn:
+        cur = conn.execute("""
+            SELECT p.shortptname, t.tblname, m.oit
+            FROM ptlist p
+            JOIN mapping m ON m.pid = p.pid
+            JOIN tblist t ON t.tid = m.tid
+            WHERE p.shortptname IS NOT NULL
+        """)
+        for short, tbl, oit in cur:
+            upper = short.strip().upper()
+            base = upper.split(".")[0].lower()
+            ctrl = allowed_bases.get(base)
+            if ctrl is not None:
+                tag_map[upper] = (tbl, int(oit), ctrl)
+    return tag_map
+
+
+def process_all(dry_run: bool, from_date: str | None, to_date: str | None) -> int:
+    """Main processing loop."""
+    print("Loading controller tag bases from register maps...", flush=True)
+    bases = load_controller_bases()
+    c3_ct = sum(1 for v in bases.values() if v == "C3")
+    c4_ct = sum(1 for v in bases.values() if v == "C4")
+    print(f"  C3: {c3_ct} base tags,  C4: {c4_ct} base tags", flush=True)
+
+    print("Building tag map from field_hist...", flush=True)
+    tag_map = build_tag_map(bases)
+    c3_tags = sum(1 for _, _, c in tag_map.values() if c == "C3")
+    c4_tags = sum(1 for _, _, c in tag_map.values() if c == "C4")
+    print(f"  Matched: C3={c3_tags}, C4={c4_tags}  (total={len(tag_map)})", flush=True)
+
+    if not tag_map:
+        print("ERROR: no tags matched — aborting", flush=True)
+        return 0
+
+    # Group by cont table
+    groups: dict[str, list[tuple[int, str, str]]] = defaultdict(list)
+    for tagname, (tbl, oit, ctrl) in tag_map.items():
+        groups[tbl].append((oit, tagname, ctrl))
+
+    print(f"  Across {len(groups)} cont tables", flush=True)
+
+    # Load tblist for ordered processing
+    with psycopg.connect(FIELD_DSN) as conn:
+        cur = conn.execute("SELECT tblname FROM tblist ORDER BY tid")
+        tbl_order = [r[0] for r in cur]
+
+    # Time range
+    t_from = (
+        datetime.strptime(from_date, "%Y-%m-%d")
+        if from_date
+        else datetime(2026, 2, 5)
+    )
+    t_to = (
+        datetime.strptime(to_date, "%Y-%m-%d") if to_date else datetime(2026, 6, 6)
+    )
+
+    total_rows = 0
+    for tbl in tbl_order:
+        col_info = groups.get(tbl)
+        if not col_info:
+            continue
+
+        # Build dynamic SELECT for mapped columns
+        oits = sorted({oit for oit, _, _ in col_info})
+        oit_map: dict[int, list[tuple[str, str]]] = defaultdict(list)
+        for oit, tag, ctrl in col_info:
+            oit_map[oit].append((tag, ctrl))
+
+        col_sel = ", ".join(f"col{c:02d}" for c in oits)
+        sql = f"SELECT dtat, {col_sel} FROM {tbl} WHERE dtat >= %s AND dtat < %s ORDER BY dtat"
+
+        print(f"\n{tbl} ({len(col_info)} cols, {len(oits)} unique oits) ...", flush=True)
+
+        with psycopg.connect(FIELD_DSN) as src_conn:
+            with src_conn.cursor(name="fetch_cursor") as src_cur:
+                src_cur.itersize = 50000
+                src_cur.execute(sql, (t_from, t_to))
+
+                batch = []
+
+                for row in src_cur:
+                    dtat, *vals = row
+                    recorded_at = dtat - timedelta(hours=9)
+                    for i, val in enumerate(vals):
+                        if val is None:
+                            continue
+                        oit = oits[i]
+                        for tagname, ctrl_id in oit_map[oit]:
+                            batch.append((tagname, str(val), recorded_at, ctrl_id))
+
+                    if len(batch) >= 50000:
+                        total_rows += _flush_batch(batch, dry_run, tbl)
+                        batch.clear()
+
+                if batch:
+                    total_rows += _flush_batch(batch, dry_run, tbl)
+
+    return total_rows
+
+
+def _flush_batch(rows: list, dry_run: bool, tbl: str) -> int:
+    """DELETE overlap + COPY batch into history_table. Returns row count."""
+    n = len(rows)
+    if dry_run:
+        print(f"  [{tbl}] {n:>8} rows (dry-run)", flush=True)
+        return n
+
+    # Collect unique tagnames + time range for DELETE
+    tagnames = list({r[0] for r in rows})
+    ts_min = min(r[2] for r in rows)
+    ts_max = max(r[2] for r in rows)
+
+    with psycopg.connect(TARGET_DSN) as tgt:
+        with tgt.cursor() as cur:
+            del_sql = """
+                DELETE FROM history_table
+                WHERE tagname = ANY(%s) AND recorded_at >= %s AND recorded_at <= %s
+            """
+            cur.execute(del_sql, (tagnames, ts_min, ts_max))
+            deleted = cur.rowcount
+
+        with tgt.cursor() as cur:
+            with cur.copy(
+                "COPY history_table (tagname, value, recorded_at, controller_id) FROM STDIN"
+            ) as copy:
+                for r in rows:
+                    copy.write_row(r)
+
+        tgt.commit()
+
+    print(f"  [{tbl}] {n:>8} inserted  ({deleted} deleted)", flush=True)
+    return n
+
+
+def main():
+    p = argparse.ArgumentParser(
+        description="Import field_hist data into hc900.history_table (C3/C4 only)"
+    )
+    p.add_argument(
+        "--dry-run", action="store_true", help="Preview only, no writes"
+    )
+    p.add_argument(
+        "--from-date", help="Start date (YYYY-MM-DD), default 2026-02-05"
+    )
+    p.add_argument(
+        "--to-date", help="End date (YYYY-MM-DD), default 2026-06-06"
+    )
+    args = p.parse_args()
+
+    t0 = time.time()
+    total = process_all(args.dry_run, args.from_date, args.to_date)
+    elapsed = time.time() - t0
+
+    print(f"\n{'=' * 60}")
+    label = "dry-run rows" if args.dry_run else "rows imported"
+    print(f"Total {label}: {total:,}")
+    print(f"Elapsed: {elapsed:.1f}s ({total / max(elapsed, 1):,.0f} rows/s)")
+    print(f"{'=' * 60}")
+
+
+if __name__ == "__main__":
+    main()

scripts/load_map_master.py

@@ -48,14 +48,19 @@ def classify(entry: dict) -> tuple[str, int | None]:
 
 def load_controller(cur, controller: str, map_path: Path, active: bool) -> int:
     data = json.loads(map_path.read_text(encoding="utf-8"))
+    # Snapshot old tagnames BEFORE DELETE so we can detect renames.
+    cur.execute(
+        "SELECT tagname FROM hc900_map_master WHERE controller_id = %s", (controller,))
+    old_names = {row[0] for row in cur.fetchall()}
+
     cur.execute("DELETE FROM hc900_map_master WHERE controller_id = %s", (controller,))
     n = 0
+    new_entries = []
     for e in data["registers"]:
         param_type, loop_no = classify(e)
         access = "R/W" if e.get("access") == "RW" else "R"
-        # Single consistent name everywhere: tagname == hc900_tag == the register tag
+        new_entries.append((e["tag"], e["tag"], e["addr"], e["type"], access,
-        # (Experion ItemName, e.g. FICQ-6101.PV) — used by the gateway, map_master and
+                            active, loop_no, param_type, controller))
-        # realtime_table alike. No case conversion.
         cur.execute(
             """INSERT INTO hc900_map_master
                  (tagname, hc900_tag, modbus_addr, data_type, access,
@@ -64,9 +69,74 @@ def load_controller(cur, controller: str, map_path: Path, active: bool) -> int:
             (e["tag"], e["tag"], e["addr"], e["type"], access,
              active, loop_no, param_type, controller))
         n += 1
+
+    # Detect renames: tagname changed between old and new map.
+    new_names = {e[0] for e in new_entries}
+    renamed = {}  # old_name -> new_name
+    for old in old_names:
+        if old not in new_names:
+            # Old tagname removed — find matching new tagname.
+            # Heuristic: same base tag, different suffix (e.g. "TI-3101" -> "TI-3101.PV")
+            for new in new_names:
+                if _is_rename_candidate(old, new):
+                    renamed[old] = new
+                    break
+
+    if renamed:
+        _migrate_history(cur, renamed, controller)
+        _migrate_realtime(cur, renamed, controller)
+        print(f"  ↻ Renamed {len(renamed)} tag(s): {', '.join(f'{k}->{v}' for k, v in renamed.items())}")
+
     return n
 
 
+def _is_rename_candidate(old: str, new: str) -> bool:
+    """Check if new could be a renamed version of old (same base, different suffix)."""
+    # Exact match → not a rename
+    if old == new:
+        return False
+    # Strip common suffixes (.PV, .SP, .OP, etc.) and compare base
+    old_base = re.sub(r'\.[A-Z_]+$', '', old)
+    new_base = re.sub(r'\.[A-Z_]+$', '', new)
+    return old_base == new_base
+
+
+def _migrate_history(cur, renamed: dict[str, str], controller: str) -> None:
+    """Rename tagname in history_table for renamed tags."""
+    if not renamed:
+        return
+    placeholders = ", ".join(["%s"] * len(renamed))
+    values = list(renamed.keys())
+    cur.execute(f"""
+        UPDATE hc900.history_table
+        SET tagname = CASE tagname
+            {"".join(f"WHEN %s THEN %s" for _ in renamed)}
+        END
+        WHERE tagname IN ({placeholders})
+          AND controller_id = %s
+    """, [*values, *renamed.values(), controller])
+    if cur.rowcount:
+        print(f"    history_table: {cur.rowcount} rows migrated")
+
+
+def _migrate_realtime(cur, renamed: dict[str, str], controller: str) -> None:
+    """Rename tagname in realtime_table for renamed tags."""
+    if not renamed:
+        return
+    placeholders = ", ".join(["%s"] * len(renamed))
+    values = list(renamed.keys())
+    cur.execute(f"""
+        UPDATE hc900.realtime_table
+        SET tagname = CASE tagname
+            {"".join(f"WHEN %s THEN %s" for _ in renamed)}
+        END
+        WHERE tagname IN ({placeholders})
+          AND controller_id = %s
+    """, [*values, *renamed.values(), controller])
+    if cur.rowcount:
+        print(f"    realtime_table: {cur.rowcount} rows migrated")
+
+
 def main():
     p = argparse.ArgumentParser(description="Load hc900_map_master from register maps")
     p.add_argument("--controller", help="controller id, e.g. C3")