writetime and TTL are not preserved if columns are not explicitly configured in cdm.properties

[icaredb]

With the following settings in cdm.properties, the writetime and TTL values are not preserved when data is migrated to target cluster. Note spark.cdm.schema.origin.column.ttl.names and spark.cdm.schema.origin.column.writetime.names are not explicitly set.

spark.cdm.schema.origin.column.ttl.automatic     true
spark.cdm.schema.origin.column.writetime.automatic  true

Although the automatic TTLCols and WriteTimestampCols are detected correctly, logging shows feature writetimeTTL is disabled:

23/11/15 01:10:08 INFO WritetimeTTL: PARAM -- Automatic TTLCols: [c3, c4]
23/11/15 01:10:08 INFO WritetimeTTL: PARAM -- Automatic WriteTimestampCols: [c3, c4]
23/11/15 01:10:08 INFO WritetimeTTL: Feature WritetimeTTL is disabled

Reproducing steps:

1. prepare on origin cluster

CREATE TABLE ks1.cdm (
    pk1 ascii,
    ck1 ascii,
    c3 ascii,
    c4 ascii,
    PRIMARY KEY (pk1, ck1)
) ;
insert into cdm(pk1,ck1,c3,c4) values ('1','x','x1','y1');
insert into cdm(pk1,ck1,c3,c4) values ('2','x','x2','y2');
update cdm set c3='x1x' where pk1='1' and ck1='x';

cqlsh> select c3,writetime(c3),c4,writetime(c4) from ks1.cdm;

 c3  | writetime(c3)    | c4 | writetime(c4)
-----+------------------+----+------------------
  x2 | 1699617875598019 | y2 | 1699617875598019
 x1x | 1699617884895474 | y1 | 1699617863817553

2. cdm.properties contents set as below:

spark.cdm.schema.origin.column.ttl.automatic     true
#spark.cdm.schema.origin.column.ttl.names       c3
spark.cdm.schema.origin.column.writetime.automatic  true
#spark.cdm.schema.origin.column.writetime.names   c3

3. writetime is not preserved after data migration on the target cluster:

cqlsh> select c3,writetime(c3),c4,writetime(c4) from ks1.cdm;

 c3  | writetime(c3)    | c4 | writetime(c4)
-----+------------------+----+------------------
  x2 | 1699869971644859 | y2 | 1699869971644859
 x1x | 1699870639789613 | y1 | 1699870639789613

[msmygit]

Implemented as part of #226