@clickzetta/cz-cli-darwin-x64 0.3.92 → 0.3.94

package/bin/skills/clickzetta-sql-migration/references/function-mapping.md

@@ -0,0 +1,194 @@
+# Function Mapping: Snowflake / Spark / Databricks → ClickZetta
+
+> Comprehensive mapping table for functions that **differ** between systems, plus a list of unsupported functions and their workarounds.
+> For the full ClickZetta function reference, refer to the official ClickZetta Lakehouse documentation.
+
+---
+
+## Conditional Functions
+
+| Snowflake | Spark / Databricks | ClickZetta | Notes |
+|---|---|---|---|
+| `IFF(cond, a, b)` | `IF(cond, a, b)` | `IF(cond, a, b)` | ClickZetta does not support `IFF` |
+| `ZEROIFNULL(x)` | — | `COALESCE(x, 0)` or `NVL(x, 0)` | |
+| `NULLIFZERO(x)` | — | `NULLIF(x, 0)` | |
+| `BOOLAND(a, b)` | — | `a AND b` | use boolean operator |
+| `BOOLOR(a, b)` | — | `a OR b` | |
+| `DECODE(...)` | `DECODE(...)` | `DECODE(...)` | ✅ all supported |
+| `NULLIF` / `COALESCE` / `NVL` | same | same | ✅ all supported |
+
+---
+
+## Date / Time Functions
+
+| Snowflake | Spark / Databricks | ClickZetta | Notes |
+|---|---|---|---|
+| `DATEADD(day, n, dt)` | `DATE_ADD(dt, n)` | `DATEADD(day, n, dt)` ✅ or `DATE_ADD(dt, n)` ✅ | both syntaxes work |
+| `DATEDIFF(day, start, end)` | `DATEDIFF(end, start)` | `DATEDIFF(day, start, end)` ✅ or `DATEDIFF(end, start)` ✅ | both supported, but **2-arg form has reversed order from Snowflake** |
+| `DATE_TRUNC('month', dt)` | `DATE_TRUNC('month', dt)` | same | ✅ identical |
+| `TO_DATE(s)` / `TO_TIMESTAMP(s)` | same | same | ✅ identical |
+| `CONVERT_TIMEZONE(tz, dt)` | `from_utc_timestamp(dt, tz)` | `FROM_UTC_TIMESTAMP(dt, tz)` / `TO_UTC_TIMESTAMP(dt, tz)` | |
+| `SYSDATE()` / `GETDATE()` | `current_timestamp()` | `CURRENT_TIMESTAMP()` or `NOW()` | both supported |
+| `TIMESTAMPADD(unit, n, dt)` | — | `dt + INTERVAL n unit` | |
+| `LAST_DAY(dt)` | `last_day(dt)` | `LAST_DAY(dt)` | ✅ identical |
+| `MONTHS_BETWEEN(d1, d2)` | `months_between(d1, d2)` | `MONTHS_BETWEEN(d1, d2)` | ✅ identical |
+| `YEAR(dt)` / `MONTH(dt)` / `DAY(dt)` | same | same | ✅ identical |
+| `DATE_PART('year', dt)` | `date_part('year', dt)` | ❌ not supported | use `EXTRACT(YEAR FROM dt)` or `YEAR(dt)` |
+| `MAKEDATE(year, dayofyear)` | — | ❌ not supported | use `MAKE_DATE(year, month, day)` |
+| `CONVERT_TZ(dt, from, to)` | — | ❌ not supported | use `FROM_UTC_TIMESTAMP` / `TO_UTC_TIMESTAMP` |
+
+---
+
+## String Functions
+
+| Snowflake | Spark / Databricks | ClickZetta | Notes |
+|---|---|---|---|
+| `CHARINDEX(sub, s)` | `instr(s, sub)` | `INSTR(s, sub)` | ⚠️ **parameter order is reversed from Snowflake** |
+| `EDITDISTANCE(s1, s2)` | `levenshtein(s1, s2)` | ❌ `LEVENSHTEIN` not supported | use Python UDF / ZettaPark |
+| `SOUNDEX(s)` | `soundex(s)` | ❌ not supported | no alternative |
+| `STRTOK(s, delim, n)` | `split(s, delim)[n-1]` | `SPLIT_PART(s, delim, n)` | |
+| `ILIKE` | `ilike` | `ILIKE` | ✅ all supported |
+| `RLIKE` / `REGEXP_LIKE` | `rlike` | `RLIKE` / `REGEXP_LIKE` | ✅ all supported |
+| `CONTAINS(s, sub)` | `contains(s, sub)` | `INSTR(s, sub) > 0` | |
+| `STARTSWITH(s, p)` | `startswith(s, p)` | `STARTSWITH(s, p)` ✅ or `s LIKE 'p%'` | both supported |
+| `ENDSWITH(s, p)` | `endswith(s, p)` | `ENDSWITH(s, p)` ✅ or `s LIKE '%p'` | both supported |
+| `INITCAP(s)` | `initcap(s)` | `INITCAP(s)` | ✅ identical |
+| `REGEXP_SUBSTR(s, p)` | `regexp_extract(s, p, 0)` | ❌ `REGEXP_SUBSTR` not supported | use `REGEXP_EXTRACT(s, '(p)')` |
+| `OVERLAY(s PLACING new FROM pos)` | `overlay(...)` | ❌ not supported | use `CONCAT(LEFT(s, pos-1), new, SUBSTR(s, pos+len))` |
+| `FORMAT(num, decimals)` | — | ❌ thousand-separator format not supported | use `ROUND` + `CAST` |
+
+---
+
+## Aggregate Functions
+
+| Snowflake | Spark / Databricks | ClickZetta | Notes |
+|---|---|---|---|
+| `LISTAGG(col, ',') WITHIN GROUP (ORDER BY col)` | `concat_ws(',', collect_list(col))` | `GROUP_CONCAT(col ORDER BY col SEPARATOR ',')` | |
+| `ARRAY_AGG(col) WITHIN GROUP (ORDER BY col)` | `array_agg(col)` (no ordering) | `ARRAY_AGG(col)` | ⚠️ `WITHIN GROUP` not supported |
+| `OBJECT_AGG(key, value)` | `map_from_entries(...)` | `MAP_AGG(key, value)` | |
+| `APPROX_COUNT_DISTINCT(col)` | `approx_count_distinct(col)` | `APPROX_COUNT_DISTINCT(col)` | ✅ identical |
+| `MEDIAN(col)` | — | `MEDIAN(col)` | ✅ identical |
+| `BITAND_AGG / BITOR_AGG / BITXOR_AGG` | — | `BIT_AND / BIT_OR / BIT_XOR` | |
+| `REGR_SLOPE / REGR_INTERCEPT` | — | ❌ not supported | manually compute via `CORR` + `STDDEV` |
+
+---
+
+## Array / Object Functions
+
+| Snowflake | Spark / Databricks | ClickZetta | Notes |
+|---|---|---|---|
+| `ARRAY_CONSTRUCT(...)` | `array(...)` | `ARRAY(...)` | |
+| `OBJECT_CONSTRUCT('k', v, ...)` | `named_struct('k', v, ...)` or `map(...)` | `named_struct('k', v, ...)` ✅ or `MAP(...)` | |
+| `ARRAY_SIZE(arr)` | `size(arr)` | `SIZE(arr)` ✅ or `ARRAY_SIZE(arr)` ✅ | both supported |
+| `ARRAY_CONTAINS(val, arr)` | `array_contains(arr, val)` | `ARRAY_CONTAINS(arr, val)` | ⚠️ **Snowflake parameter order reversed** |
+| `OBJECT_KEYS(obj)` | `map_keys(map)` | `MAP_KEYS(map)` | |
+| `FLATTEN(arr)` | `flatten(arr)` | `FLATTEN(arr)` | ✅ for array of arrays |
+| `LATERAL FLATTEN(input => arr)` | `LATERAL VIEW EXPLODE(arr)` | `LATERAL VIEW EXPLODE(arr)` | ⚠️ Snowflake → Hive-style syntax change |
+| `STRUCT(1 AS id, 'a' AS name)` (Spark) | same | `named_struct('id', 1, 'name', 'a')` | ⚠️ ClickZetta `STRUCT` does not accept `AS` for named fields |
+| `TO_ARRAY(expr)` | — | ❌ not supported | use `ARRAY(expr)` or `CAST(... AS ARRAY<T>)` |
+| `MAP_FROM_ZIP(keys, values)` | — | ❌ not supported | use `MAP_FROM_ARRAYS(keys, values)` |
+
+ClickZetta supports higher-order functions (Spark style) which Snowflake does not:
+
+```sql
+SELECT TRANSFORM(skills, x -> UPPER(x)) FROM emp;
+SELECT FILTER(scores, x -> x > 90) FROM students;
+SELECT EXISTS(scores, x -> x > 100) FROM students;
+SELECT FORALL(scores, x -> x >= 0) FROM students;
+SELECT ZIP_WITH(a, b, (x, y) -> x + y) FROM t;
+```
+
+`AGGREGATE` / `REDUCE` (Spark names) are not supported — use `ARRAY_AGG` + aggregate functions instead.
+
+---
+
+## JSON / Semi-structured Access
+
+```sql
+-- Snowflake (colon syntax + double-colon cast)
+SELECT data:address:city AS city FROM users;
+SELECT data:age::INT AS age FROM users;
+SELECT data:phoneNumbers[0]:number FROM users;
+
+-- ClickZetta (bracket syntax)
+SELECT data['address']['city'] AS city FROM users;
+SELECT CAST(data['age'] AS INT) AS age FROM users;
+SELECT data['phoneNumbers'][0]['number'] FROM users;
+
+-- ClickZetta also accepts :: cast operator
+SELECT data['amount']::DOUBLE AS amount FROM orders;
+```
+
+| Snowflake | ClickZetta |
+|---|---|
+| `data:key` | `data['key']` |
+| `data[0]:key` | `data[0]['key']` |
+| `data:key::TYPE` | `CAST(data['key'] AS TYPE)` or `data['key']::TYPE` |
+| `PARSE_JSON(s)` | `PARSE_JSON(s)` ✅ identical |
+| `TO_VARIANT(x)` | `PARSE_JSON(TO_JSON(x))` |
+| `TO_JSON(x)` | `TO_JSON(x)` ✅ identical |
+| `IS_NULL_VALUE(json:key)` | `data['key'] IS NULL` |
+
+---
+
+## System / Context Functions
+
+| Snowflake | Spark / Databricks | ClickZetta | Notes |
+|---|---|---|---|
+| `CURRENT_DATABASE()` | `current_database()` | `CURRENT_WORKSPACE()` | concept rename |
+| `CURRENT_WAREHOUSE()` | — | `CURRENT_VCLUSTER()` | concept rename |
+| `CURRENT_ROLE()` | `current_user()` | `CURRENT_USER()` | no role function |
+| `CURRENT_SCHEMA()` | `current_database()` | `CURRENT_SCHEMA()` | ✅ |
+| — | — | `CURRENT_INSTANCE_ID()` | ClickZetta-specific |
+
+---
+
+## Type Conversion Functions
+
+| Snowflake | Spark / Databricks | ClickZetta | Notes |
+|---|---|---|---|
+| `TRY_TO_NUMBER(s)` / `TRY_TO_DATE(s)` | `try_cast(s AS ...)` | `TRY_CAST(s AS ...)` | |
+| `TO_VARIANT(x)` | — | `PARSE_JSON(TO_JSON(x))` | |
+| `CAST(...)` / `::TYPE` | `CAST(...)` / `::TYPE` | `CAST(...)` / `::TYPE` | ✅ all supported |
+
+---
+
+## Functions with No Direct ClickZetta Equivalent
+
+| Function | Source | Workaround |
+|---|---|---|
+| `SOUNDEX(s)` | Snowflake | None |
+| `EDITDISTANCE` / `LEVENSHTEIN` | Snowflake / Spark | Python UDF |
+| `JSON_ARRAY_LENGTH` | various | `SIZE(CAST(json AS ARRAY<STRING>))` |
+| `JSON_OBJECT_KEYS` | various | manually parse |
+| `REGEXP_SUBSTR` | Snowflake | `REGEXP_EXTRACT(s, '(p)')` |
+| `GENERATE_SERIES(s, e)` / `RANGE(n)` | various | `EXPLODE(SEQUENCE(s, e))` |
+| `TABLESAMPLE (n PERCENT)` | various | `ORDER BY RAND() LIMIT n` |
+| `ST_*` geospatial functions | various | None — geospatial not supported |
+| `TO_IPV4` / IP address functions | various | None |
+| `HLL_APPROX` | various | `APPROX_COUNT_DISTINCT(col)` |
+| `BITAND(a, b)` / `BITOR(a, b)` / `BITXOR(a, b)` | various | bitwise operators `&` / `\|` / `^` |
+| `INITCAP(s)` (in versions that miss it) | — | `CONCAT(UPPER(SUBSTR(s,1,1)), LOWER(SUBSTR(s,2)))` |
+| `SQUARE(x)` | Snowflake | `POWER(x, 2)` |
+| `HAVERSINE(...)` | Snowflake | None |
+| `WIDTH_BUCKET(...)` | Snowflake | None |
+| `FACTORIAL(n)` | various | `EXP(SUM(LN(generate)))` over a sequence |
+| `BIN(x)` | various | `CONV(x, 10, 2)` |
+
+---
+
+## Vector Functions (ClickZetta-Specific)
+
+ClickZetta has native vector functions for similarity search, which Snowflake/Spark do not provide:
+
+```sql
+L2_DISTANCE(v1, v2)             -- Euclidean distance
+COSINE_DISTANCE(v1, v2)         -- Cosine distance
+DOT_PRODUCT(v1, v2)             -- Dot product
+HAMMING_DISTANCE(v1, v2)        -- Hamming distance (binary)
+JACCARD_DISTANCE(v1, v2)        -- Jaccard distance
+BINARY_QUANTIZE(v)              -- float vector → binary
+VECTOR(v1, v2, ...)             -- construct vector
+```
+
+If migrating from Snowflake Cortex Search or Databricks Vector Search, redesign around these primitives + the `VECTOR INDEX` (see ClickZetta Lakehouse documentation).

package/bin/skills/clickzetta-sql-migration/references/functions-reference.md

@@ -0,0 +1,372 @@
+# Functions Complete Reference
+
+> With Snowflake / Spark SQL difference annotations
+
+---
+
+## Numeric Functions
+
+```sql
+ABS(x)                          -- absolute value
+CEIL(x) / CEILING(x)            -- round up
+FLOOR(x)                        -- round down
+ROUND(x, d)                     -- round to d decimal places
+TRUNCATE(x, d)                  -- truncate to d decimal places
+MOD(x, y) / x % y               -- modulo
+POWER(x, y) / POW(x, y)         -- exponentiation
+SQRT(x)                         -- square root
+EXP(x)                          -- e^x
+LN(x) / LOG(x)                  -- natural logarithm
+LOG(base, x)                    -- logarithm with specified base
+LOG2(x) / LOG10(x)              -- base-2/base-10 logarithm
+SIGN(x)                         -- sign (-1/0/1)
+GREATEST(a, b, c, ...)          -- maximum value
+LEAST(a, b, c, ...)             -- minimum value
+RANDOM() / RAND()               -- random number 0-1
+PI()                            -- π
+SIN(x) / COS(x) / TAN(x)       -- trigonometric functions
+ASIN(x) / ACOS(x) / ATAN(x)    -- inverse trigonometric functions
+ATAN2(y, x)                     -- arctangent
+DEGREES(x) / RADIANS(x)        -- degree/radian conversion
+-- ⚠️ FACTORIAL not supported, use EXP(SUM(LN(n))) instead
+-- ⚠️ BIN(x) not supported, use CONV(x, 10, 2) instead
+HEX(x)                          -- convert to hexadecimal string
+UNHEX(s)                        -- hexadecimal to string
+CONV(x, from_base, to_base)     -- base conversion (e.g., CONV(10,10,2) gives '1010')
+```
+
+**Differences from Snowflake:**
+- Snowflake `SQUARE(x)` → ClickZetta `POWER(x, 2)`
+- Snowflake `HAVERSINE(lat1, lon1, lat2, lon2)` → ClickZetta not supported
+- Snowflake `WIDTH_BUCKET` → ClickZetta not supported
+
+---
+
+## String Functions
+
+```sql
+-- Basic operations
+LENGTH(s) / CHAR_LENGTH(s)      -- character length
+OCTET_LENGTH(s)                 -- byte length
+UPPER(s) / LOWER(s)             -- case conversion
+INITCAP(s)                      -- capitalize first letter
+TRIM(s) / LTRIM(s) / RTRIM(s)  -- trim whitespace
+TRIM(BOTH 'x' FROM s)           -- trim specified character
+LPAD(s, n, pad) / RPAD(s, n, pad)  -- padding
+REPEAT(s, n)                    -- repeat
+REVERSE(s)                      -- reverse
+SPACE(n)                        -- n spaces
+
+-- Concatenation
+CONCAT(s1, s2, ...)             -- concatenate (NULL propagates)
+CONCAT_WS(sep, s1, s2, ...)     -- concatenate with separator (skips NULL)
+s1 || s2                        -- concatenation operator
+
+-- Substring
+SUBSTR(s, pos) / SUBSTRING(s, pos)
+SUBSTR(s, pos, len) / SUBSTRING(s, pos, len)
+LEFT(s, n) / RIGHT(s, n)
+MID(s, pos, len)                -- same as SUBSTR
+
+-- Search
+INSTR(s, substr)                -- find position (1-based, 0 means not found)
+LOCATE(substr, s)               -- same as INSTR, different parameter order
+LOCATE(substr, s, pos)          -- search from pos
+POSITION(substr IN s)           -- ✅ supported, returns substring position (1-based)
+FIND_IN_SET(s, list)            -- find in comma-separated list
+
+-- Replace
+REPLACE(s, old, new)            -- replace all occurrences
+TRANSLATE(s, from_chars, to_chars)  -- character-level replacement
+-- ⚠️ OVERLAY syntax not supported, use CONCAT(LEFT(s,pos-1), new, SUBSTR(s,pos+len)) instead
+
+-- Regex
+REGEXP_EXTRACT(s, pattern, group)   -- extract matching group
+REGEXP_EXTRACT_ALL(s, pattern)      -- extract all matches
+REGEXP_REPLACE(s, pattern, repl)    -- regex replace
+REGEXP_LIKE(s, pattern)             -- regex match (returns boolean)
+RLIKE(s, pattern)                   -- same as REGEXP_LIKE
+s RLIKE pattern                     -- operator form
+REGEXP_COUNT(s, pattern)            -- match count
+REGEXP_SUBSTR(s, pattern)           -- extract first match
+
+-- Split
+SPLIT(s, delimiter)             -- split by delimiter, returns ARRAY
+SPLIT_PART(s, delimiter, n)     -- get nth split part (1-based)
+
+-- Formatting
+FORMAT_STRING(fmt, args...)     -- printf style (e.g., FORMAT_STRING('%d items', 5) → '5 items')
+-- ⚠️ FORMAT(number, decimals) number thousand-separator formatting not supported, use ROUND + CAST instead
+
+-- Encoding
+BASE64(s) / UNBASE64(s)         -- Base64 encode/decode
+MD5(s)                          -- MD5 hash
+SHA1(s) / SHA2(s, bits)         -- SHA hash
+CRC32(s)                        -- CRC32
+ENCODE(s, charset) / DECODE(s, charset)  -- charset encode/decode
+
+-- Other
+ASCII(s)                        -- ASCII code of first character
+CHAR(n)                         -- ASCII code to character
+-- ⚠️ SOUNDEX not supported
+-- ⚠️ LEVENSHTEIN not supported, use Python UDF or ZettaPark instead
+HAMMING_DISTANCE(s1, s2)        -- Hamming distance (strings)
+```
+
+**Differences from Snowflake:**
+- Snowflake `CHARINDEX(substr, s)` → ClickZetta `INSTR(s, substr)` or `LOCATE(substr, s)` (different parameter order!)
+- Snowflake `EDITDISTANCE(s1, s2)` → ClickZetta does not support LEVENSHTEIN, use Python UDF
+- Snowflake `STRTOK(s, delim, n)` → ClickZetta `SPLIT_PART(s, delim, n)`
+- Snowflake `ILIKE(s, pattern)` → ClickZetta `ILIKE` ✅ also supported!
+- Snowflake `CONTAINS(s, substr)` → ClickZetta `INSTR(s, substr) > 0`
+- Snowflake `STARTSWITH(s, prefix)` → ClickZetta `s LIKE 'prefix%'` or `STARTSWITH(s, prefix)`
+- Snowflake `ENDSWITH(s, suffix)` → ClickZetta `s LIKE '%suffix'` or `ENDSWITH(s, suffix)`
+
+---
+
+## Date/Time Functions
+
+```sql
+-- Get current time
+CURRENT_DATE()                  -- current date
+CURRENT_TIMESTAMP() / NOW()     -- current timestamp (with timezone)
+CURRENT_TIME()                  -- current time
+LOCALTIMESTAMP()                -- local timestamp
+
+-- Extract parts
+YEAR(dt) / MONTH(dt) / DAY(dt)
+HOUR(dt) / MINUTE(dt) / SECOND(dt)
+DAYOFWEEK(dt)                   -- 1=Sunday, 7=Saturday
+DAYOFMONTH(dt)                  -- same as DAY
+DAYOFYEAR(dt)                   -- day of year
+WEEKOFYEAR(dt)                  -- week of year
+QUARTER(dt)                     -- quarter (1-4)
+EXTRACT(YEAR FROM dt)           -- standard SQL extraction
+-- ⚠️ DATE_PART('year', dt) not supported, use EXTRACT or YEAR(dt) instead
+
+-- Date arithmetic
+DATE_ADD(dt, n)                 -- add n days
+DATE_SUB(dt, n)                 -- subtract n days
+dt + INTERVAL n DAY             -- add n days (standard SQL)
+dt - INTERVAL n DAY             -- subtract n days
+dt + INTERVAL '1-2' YEAR TO MONTH  -- add 1 year 2 months
+ADDDATE(dt, n)                  -- same as DATE_ADD
+SUBDATE(dt, n)                  -- same as DATE_SUB
+ADD_MONTHS(dt, n)               -- add n months
+MONTHS_BETWEEN(dt1, dt2)        -- month difference
+
+-- Date difference
+DATEDIFF(end_dt, start_dt)      -- two-parameter form: returns day difference (end first)
+DATEDIFF(unit, start_dt, end_dt) -- three-parameter form: specify unit (day/hour/month etc.), Snowflake-compatible
+TIMESTAMPDIFF(unit, dt1, dt2)   -- difference in specified unit
+
+-- Truncation
+DATE_TRUNC('year', dt)          -- truncate to year
+DATE_TRUNC('month', dt)         -- truncate to month
+DATE_TRUNC('day', dt)           -- truncate to day
+DATE_TRUNC('hour', dt)          -- truncate to hour
+DATE_TRUNC('week', dt)          -- truncate to week (Monday)
+TRUNC(dt, 'MM')                 -- Oracle-style truncation
+
+-- Formatting
+DATE_FORMAT(dt, 'yyyy-MM-dd')   -- format to string
+DATE_FORMAT(dt, 'yyyy-MM-dd HH:mm:ss')
+TO_CHAR(dt, 'YYYY-MM-DD')       -- same as DATE_FORMAT
+
+-- Conversion
+TO_DATE('2024-01-01')           -- string to date
+TO_DATE('2024-01-01', 'yyyy-MM-dd')
+TO_TIMESTAMP('2024-01-01 12:00:00')
+TO_TIMESTAMP('2024-01-01', 'yyyy-MM-dd')
+CAST('2024-01-01' AS DATE)
+CAST('2024-01-01 12:00:00' AS TIMESTAMP)
+FROM_UNIXTIME(unix_ts)          -- Unix timestamp to timestamp
+FROM_UNIXTIME(unix_ts, fmt)     -- to formatted string
+UNIX_TIMESTAMP()                -- current Unix timestamp
+UNIX_TIMESTAMP(dt)              -- date to Unix timestamp
+UNIX_TIMESTAMP(s, fmt)          -- string to Unix timestamp
+
+-- Other
+LAST_DAY(dt)                    -- last day of month
+NEXT_DAY(dt, 'Monday')          -- next specified day of week
+MAKE_DATE(year, month, day)     -- construct date (note: MAKE_DATE not MAKEDATE)
+ADD_MONTHS(dt, n)               -- add n months
+MONTHS_BETWEEN(dt1, dt2)        -- month difference
+TIMESTAMPDIFF(unit, dt1, dt2)   -- difference in specified unit (e.g., TIMESTAMPDIFF(MONTH, ...))
+FROM_UTC_TIMESTAMP(ts, tz)      -- UTC to specified timezone
+TO_UTC_TIMESTAMP(ts, tz)        -- specified timezone to UTC
+-- ⚠️ CONVERT_TZ(dt, from_tz, to_tz) not supported, use FROM_UTC_TIMESTAMP/TO_UTC_TIMESTAMP instead
+-- ⚠️ MAKEDATE(year, dayofyear) not supported, use MAKE_DATE(year, month, day) instead
+-- ⚠️ MAKETIME / PERIOD_ADD / PERIOD_DIFF not supported
+```
+
+**Differences from Snowflake:**
+- Snowflake `DATEADD(day, n, dt)` → ClickZetta `DATEADD(day, n, dt)` ✅ also supported; or use `DATE_ADD(dt, n)` / `dt + INTERVAL n DAY`
+- Snowflake `DATEDIFF(day, start, end)` → ClickZetta `DATEDIFF(day, start, end)` ✅ three-parameter form also supported; or use `DATEDIFF(end, start)` two-parameter form (returns days)
+- Snowflake `DATE_TRUNC('day', dt)` → ClickZetta same
+- Snowflake `TO_DATE(s)` → ClickZetta same
+- Snowflake `CONVERT_TIMEZONE(from, to, ts)` → ClickZetta `FROM_UTC_TIMESTAMP` / `TO_UTC_TIMESTAMP`
+- Snowflake `CONVERT_TIMEZONE(tz, dt)` → ClickZetta `CONVERT_TZ(dt, from_tz, to_tz)`
+- Snowflake `SYSDATE()` / `GETDATE()` → ClickZetta `CURRENT_TIMESTAMP()` / `NOW()`
+- Snowflake `TIMESTAMPADD(unit, n, dt)` → ClickZetta `dt + INTERVAL n unit`
+
+**Differences from Spark SQL:**
+- Most functions are the same; ClickZetta is compatible with Spark date functions
+
+---
+
+## Conditional Functions
+
+```sql
+-- IF
+IF(condition, true_val, false_val)
+
+-- CASE WHEN
+CASE WHEN cond1 THEN val1
+     WHEN cond2 THEN val2
+     ELSE default_val
+END
+
+-- Simple CASE
+CASE status
+    WHEN 'A' THEN 'Active'
+    WHEN 'I' THEN 'Inactive'
+    ELSE 'Unknown'
+END
+
+-- NULL handling
+COALESCE(a, b, c)               -- first non-NULL value
+NVL(a, b)                       -- return b if a is NULL (same as IFNULL)
+IFNULL(a, b)                    -- same as NVL
+NULLIF(a, b)                    -- return NULL if a=b, otherwise return a
+NVL2(a, b, c)                   -- return b if a is not NULL, otherwise c
+ISNULL(a)                       -- is NULL (returns boolean)
+ISNOTNULL(a)                    -- is not NULL
+
+-- DECODE (Oracle/Hive style)
+DECODE(expr, val1, res1, val2, res2, ..., default)
+
+-- Type checking
+TYPEOF(expr)                    -- returns type name as string
+```
+
+**Differences from Snowflake:**
+- Snowflake `IFF(cond, a, b)` → ClickZetta `IF(cond, a, b)`
+- Snowflake `ZEROIFNULL(x)` → ClickZetta `COALESCE(x, 0)` or `NVL(x, 0)`
+- Snowflake `NULLIFZERO(x)` → ClickZetta `NULLIF(x, 0)`
+- Snowflake `BOOLAND(a, b)` / `BOOLOR(a, b)` → ClickZetta `a AND b` / `a OR b`
+
+---
+
+## Aggregate Functions
+
+```sql
+-- Basic aggregation
+COUNT(*) / COUNT(col) / COUNT(DISTINCT col)
+SUM(col) / AVG(col) / MAX(col) / MIN(col)
+STDDEV(col) / STDDEV_POP(col) / STDDEV_SAMP(col)
+VARIANCE(col) / VAR_POP(col) / VAR_SAMP(col)
+
+-- Boolean aggregation
+BOOL_OR(cond)                   -- any one is true
+BOOL_AND(cond)                  -- all are true
+EVERY(cond)                     -- same as BOOL_AND
+
+-- String aggregation
+GROUP_CONCAT(col ORDER BY col SEPARATOR ',')   -- replaces Snowflake LISTAGG
+GROUP_CONCAT(DISTINCT col SEPARATOR ',')
+
+-- Array aggregation
+ARRAY_AGG(col)                  -- collect into array (includes NULL)
+COLLECT_LIST(col)               -- same as ARRAY_AGG
+COLLECT_SET(col)                -- collect deduplicated
+
+-- Approximate aggregation
+APPROX_COUNT_DISTINCT(col)      -- approximate distinct count (HyperLogLog)
+APPROX_PERCENTILE(col, p)       -- approximate percentile
+
+-- Statistical aggregation
+CORR(x, y)                      -- correlation coefficient
+COVAR_POP(x, y) / COVAR_SAMP(x, y)  -- covariance
+-- ⚠️ REGR_SLOPE / REGR_INTERCEPT not supported
+-- Alternative: CORR(y,x) * STDDEV(y) / STDDEV(x) to calculate slope
+
+-- Ordered set aggregation
+PERCENTILE(col, p)              -- exact percentile
+PERCENTILE_APPROX(col, p)       -- approximate percentile
+MEDIAN(col)                     -- median
+```
+
+**Differences from Snowflake:**
+- Snowflake `LISTAGG(col, ',') WITHIN GROUP (ORDER BY col)` → ClickZetta `GROUP_CONCAT(col ORDER BY col SEPARATOR ',')`
+- Snowflake `ARRAY_AGG(col) WITHIN GROUP (ORDER BY col)` → ClickZetta `ARRAY_AGG(col)` does not support WITHIN GROUP
+- Snowflake `OBJECT_AGG(key, value)` → ClickZetta `MAP_AGG(key, value)`
+- Snowflake `BITAND_AGG / BITOR_AGG / BITXOR_AGG` → ClickZetta `BIT_AND / BIT_OR / BIT_XOR`
+
+---
+
+## Type Conversion Functions
+
+```sql
+-- Explicit conversion
+CAST(expr AS target_type)
+expr::target_type               -- shorthand syntax
+
+-- Safe conversion (returns NULL on failure instead of error)
+TRY_CAST(expr AS target_type)
+
+-- String conversion
+TO_NUMBER(s) / TO_DECIMAL(s)
+TO_DOUBLE(s)
+TO_BOOLEAN(s)                   -- 'true'/'false'/'1'/'0'
+
+-- Examples
+CAST('123' AS INT)
+CAST(123 AS STRING)
+CAST('2024-01-01' AS DATE)
+CAST('[1,2,3]' AS VECTOR(3))    -- string to vector
+TRY_CAST('abc' AS INT)          -- returns NULL
+```
+
+**Differences from Snowflake:**
+- Snowflake `TRY_TO_NUMBER / TRY_TO_DATE` → ClickZetta `TRY_CAST`
+- Snowflake `TO_VARIANT(x)` → ClickZetta `PARSE_JSON(TO_JSON(x))`
+
+---
+
+## System/Context Functions
+
+```sql
+CURRENT_USER()                  -- current username
+CURRENT_WORKSPACE()             -- current workspace
+CURRENT_SCHEMA()                -- current schema
+CURRENT_VCLUSTER()              -- current compute cluster
+CURRENT_INSTANCE_ID()           -- current instance ID
+VERSION()                       -- version information
+```
+
+**Differences from Snowflake:**
+- Snowflake `CURRENT_DATABASE()` → ClickZetta `CURRENT_WORKSPACE()`
+- Snowflake `CURRENT_WAREHOUSE()` → ClickZetta `CURRENT_VCLUSTER()`
+- Snowflake `CURRENT_ROLE()` → ClickZetta has no direct equivalent
+
+---
+
+## Vector Functions
+
+```sql
+-- Distance calculation
+L2_DISTANCE(v1, v2)             -- Euclidean distance (smaller = more similar)
+COSINE_DISTANCE(v1, v2)         -- Cosine distance (smaller = more similar)
+DOT_PRODUCT(v1, v2)             -- Dot product (larger = more similar, requires normalization)
+HAMMING_DISTANCE(v1, v2)        -- Hamming distance (binary vectors)
+JACCARD_DISTANCE(v1, v2)        -- Jaccard distance
+
+-- Vector operations
+BINARY_QUANTIZE(v)              -- binarize float vector
+VECTOR(v1, v2, ...)             -- build vector
+
+-- Build vector
+SELECT VECTOR(0.1, 0.2, 0.3, 0.4);
+SELECT CAST('[0.1, 0.2, 0.3]' AS VECTOR(3));
+```